UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Stratosphere troposphere exchange and its influence on surface ozone concentrations in the lower Fraser… Bovis, Paul E. 2003

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
831-ubc_2003-0580.pdf [ 6.1MB ]
Metadata
JSON: 831-1.0052336.json
JSON-LD: 831-1.0052336-ld.json
RDF/XML (Pretty): 831-1.0052336-rdf.xml
RDF/JSON: 831-1.0052336-rdf.json
Turtle: 831-1.0052336-turtle.txt
N-Triples: 831-1.0052336-rdf-ntriples.txt
Original Record: 831-1.0052336-source.json
Full Text
831-1.0052336-fulltext.txt
Citation
831-1.0052336.ris

Full Text

S T R A T O S P H E R E O N  S U R F A C E  T R O P O S P H E R E E X C H A N G E A N D  O Z O N E  C O N C E N T R A T I O N S  I N  T H E  I T S  I N F L U E N C E  L O W E R  V A L L E Y  by Paul E. Bovis B S c , University of British Columbia, 2001 A THESIS S U B M I T T E D IN P A R T I A L F U L F I L M E N T OF T H E R E Q U I R E M E N T S FOR T H E D E G R E E OF M A S T E R OF S C I E N C E in T H E F A C U L T Y OF G R A D U A T E STUDIES (Atmospheric Science) We accept this thesis as conforming to the required standard  T H E U N I V E R S I T Y OF BRITISH C O L U M B I A August 28, 2003 © Paul E. Bovis, 2003  F R A S E R  In presenting this thesis in partial fulfillment for the requirements for an advanced degree at the University of British Columbia, I agree that the library shall make it freely available for reference and study. I further agree that permission for extensive copying of this thesis for scholarly purposes may be granted by the head of my department or by his or her representatives. It is understood that copying and publication of this thesis for financial gain shall not be allowed without my written permission.  Atmospheric Science The University of British Columbia Vancouver, Canada  Date  ,10^3  Abstract 7  B e activity concentration data collected at the University of British Columbia and  ozone data collected throughout the Lower Fraser Valley ( L F V ) , British Columbia, Canada from A p r i l 1996 - M a r c h 2001 were used to investigate the frequency, magnitude and source regions of possible stratosphere-troposphere exchange (STE) episodes that may have influenced surface ozone concentrations in the L F V . W i t h the aid of synoptic and back trajectory analysis, it was determined that the bulk of the intruded stratospheric ozone, which resulted in elevated surface ozone concentrations in the L F V , originated within intense Eastern Asian/Western Pacific cutoff low pressure systems and troughs. This stratospheric air was then transported by the synoptic scale flow to the L F V region where it descended to the surface via synoptic scale high pressure systems. During these S T E episodes, it was estimated that stratospheric ozone concentrations of up to 40 ppb (depending on the season and time of day) were contributing to the total surface ozone concentrations in the L F V .  Since anthropogenic ozone production is already a major  concern of air quality management programs in the L F V , this stratospheric contribution cannot be ignored when decisions are being made with regards to surface ozone reduction initiatives in this region.  ii  Table of Contents Abstract Table of Contents List of Tables List of Figures Acknowledgements 1 Introduction  ii iii v vi ix 1  1.1 In-Situ Aircraft Studies 1.2 High Altitude Studies 1.3 Sea Level Studies 1.4 Research Goals  4 5 6 8  2 Data  2.1 Be in the Atmosphere 2.1.1 Radionuclide Measurements as Part of the CTBT 2.1.2 Be Measurements in Vancouver, Canada 2.2 Ozone 2.2.1 Ozone Formation in the Stratosphere 2.2.2 Ozone Formation in the Troposphere 2.2.3 The GVRD Air Quality Monitoring Network 2.3 Temperature and Precipitation 2.4 Satellite Data 7  7  3 Data Analysis  3.1 80* Percentile Algorithm 3.1.1 80 Percentile Algorithm Using All Data 3.1.2 80 Percentile Algorithm Excluding June to August 3.1.3 80 Percentile Algorithm Excluding May to September 3.1.4 Disadvantages of the 80 Percentile Algorithm 3.2 Be Only Algorithm 3.3 Summary th  th  th  th  7  4 Trajectories and Case Studies  4.1 Previous Research 4.2 Trajectory Models and Procedures Used in the Current Research 4.2.1 HYSPLIT 4.2.2 CANERM 4.2.3 Trajectory Model Procedure 4.3 Case Studies 4.3.1 Case Study #1-28 October 1996-31 October 1996 4.3.1.1 Synoptic Overview 4.3.1.2 Surface 0 and Be Concentrations and Temperature in the LFV 4.3.1.3 Back Trajectory Analysis 4.3.2 Case Study #2-20 December 1996-01 January 1997 4.3.2.1 Synoptic Overview 4.3.2.2 Surface 0 and Be Concentrations and Temperature in the LFV 4.3.2.3 Back Trajectory Analysis 4.3.3 Case Study #3 - 22 January 1997 - 29 January 1997 4.3.3.1 Synoptic Overview 4.3.3.2 Surface 0 and Be Concentrations and Temperature in the LFV 4.3.3.3 Back Trajectory Analysis 4.3.4 Case Study #4-05 January 2001 - 11 January 2001 7  3  7  3  7  3  iii  10  10 13 14 17 17 19 19 25 25  25  26 27 33 35 35 36 47  50  51 54 54 54 55 55 57 57 58 63 65 65 70 71 73 73 79 80 83  4.3.4.1 Synoptic Overview 4.3.4.2 Surface 0 and Be Concentrations and Temperature in the LFV 4.3.4.3 Back Trajectory Analysis 7  3  5 Discussion  83 91 92 96  5.1 Possible Source Regions of Ozone 5.2 General Trajectory Error - A Brief Discussion 5.3 CANERM and HYSPLIT Model Comparison 5.4 Precipitation and its Influence on Be Concentrations 7  6 Conclusion  96 98 100 106 108  6.1 Future Research  111  7 References  113  iv  List of Tables Table 1 Table 2 Table 3 Table 4 Table 5 -  Episodes based on the 80th percentile detection algorithm (all data). Episodes based on the 80th percentile detection algorithm (excluding June August). Episodes based on the 80th percentile detection algorithm (excluding M a y September). Episodes based on the B e detection algorithm. G V R D A i r Quality Monitoring stations used in the hourly ozone concentration graphs in Figure 16. 7  List of Figures Figure 1 -  A modeled north-south section through the atmosphere showing the surfaces of constant B e production. The modeled vertical concentration profiles of B e at 45°N for settling velocities (1) v = 0 ms" , (2) rj = 0.02 ms" and (3) v = 1 ms" . M a p of the Comprehensive Nuclear Test B a n Treaty Internaional Montoring System radionuclide network. Daily B e activity concentrations measured at the C T B T O I M S station C A 0 0 2 from 9 A p r i l 1996- 12 March 2002. Cumulative distribution of the B e activity concentrations at the C T B T O I M S station C A 0 0 2 during the period 9 A p r i l 1 9 9 6 - 12 M a r c h 2002. Monthly average B e concentration at the C T B T O I M S station C A 0 0 2 during the period 9 A p r i l 1996- 12 March 2002. Average number of days in a month in which B e activity concentration exceeds the 80th percentile at the C T B T O I M S station C A 0 0 2 during the period 9 A p r i l 1996- 12 March 2002. Chemical reactions of Chapman's mechanism, a static pure oxygen photochemical steady state model. A generic sequence of reactions for the loss of odd oxygen by way of a catalytic cycle. 7  Figure 2 -  7  1  s  Figure 3 Figure 4 Figure 5 Figure 6 Figure 7 -  Figure 8 Figure 9 Figure 10 -  Figure 11 Figure 12 -  Figure 13 -  1  Figure 15 -  s  7  7  7  7  Observed (solid line) and calculated (dashed line) vertical concentration profiles of ozone as a function of latitude in the Northern Hemisphere for mid-spring. Greater Vancouver Regional District Continuous A i r Quality Monitoring Network. Composite diumal ozone concentration graphs for Downtown Vancouver (T001), Kitsilano (T002), Chilliwack Airport (TO 12), Surrey East (TO 15), Richmond (TO 17), Langley Central (T027), Hope Airport (T029) and Port Coquitlam North (T032). B e activity concentration, average daily maximum temperature, average daily maximum ozone concentration and coefficient of variance of ozone for 3 M a y 1996 - 22 M a y 1996, 13 July 1997 - 25 July 1997, 13 A p r i l 2000 - 25 A p r i l 2000 and 11 February 2001 - 20 February 2001 from the 8 0 percentile detection algorithm (all data). Average number of days in a month in which the average daily maximum temperature exceeded 20°C in the Lower Fraser V a l l e y during the period 10 A p r i l 1 9 9 6 - 3 1 December 2001. B e activity concentration, average daily maximum temperature, average daily maximum ozone concentration and coefficient of variance of ozone for 25 October 1996 - 5 November 1996, 21 December 1996- 1 January 1997, 22 January 1 9 9 7 - 1 February 1997 and 4 January 2001 - 12 January 2001 from the B e only detection algorithm.  7  th  Figure 14 -  1  s  7  7  vi  Figure 16 -  Figure 17 -  Figure 18 -  Figure 19 -  Hourly ozone concentrations measured at Kitsilano, Richmond South, Port Coquitlam North and Downtown Vancouver urban air quality monitoring stations and Hope Airport, Chilliwack Airport, Surrey East and Langley Central rural air quality monitoring stations in the Greater Vancouver Regional District for 28 October 1996 - 1 November 1996, 20 December 1996 - 7 January 1997, 21 January 197 - 28 January 1997 and 6 January 2001 - 9 January 2001. Hourly ozone concentrations measured Hope Airport, Chilliwack Airport, Surrey East and Langley Central rural air quality monitoring stations in the Greater Vancouver Regional District for 20 July 1996 - 29 July 1996. Hourly ozone concentrations measured Hope Airport, Chilliwack Airport, Surrey East and Langley Central rural air quality monitoring stations in the Greater Vancouver Regional District for 2 November 1998 - 11 November 1998. Total number of days in a month in which both the daily B e activity concentration and the average daily maximum ozone concentration exceeded the 8 0 percentile (all data) at the C T B T O I M S station C A 0 0 2 during the period 9 A p r i l 1996- 12 March 2002. Total number of days in a month in which both the daily B e activity concentration and the average daily maximum ozone concentration exceeded the 8 0 percentile (excluding June - August) at the C T B T O I M S station C A 0 0 2 during the period 9 A p r i l 1 9 9 6 - 12 M a r c h 2002. Total number of days in a month in which both the daily B e activity concentration and the average daily maximum ozone concentration exceeded the 8 0 percentile (excluding M a y - September) at the C T B T O I M S station C A 0 0 2 during the period 9 A p r i l 1996 - 12 M a r c h 2002. Total number of days in a month in which the daily B e activity concentration exceeded 6000 uBqm" at the C T B T O I M S station C A 0 0 2 during the period 9 A p r i l 1996- 12 March 2002. 500 mb geopotential chart for 25 October 1996 at 1800 U T C . 250 mb geopotential chart for 25 October 1996 at 1800 U T C . Surface isobaric chart for 28 October 1996 at 0000 U T C . Three dimensional back trajectories arriving over the L F V area on 31 October 1996 at 0600 U T C using the C A N E R M model. 500 mb geopotential chart for 23 December 1996 at 0600 U T C . 250 mb geopotential chart for 23 December 1996 at 0600 U T C . Surface isobaric chart for 28 December 1996 at 1800 U T C . Three dimensional back trajectories arriving over the L F V area on 26 December 1996 at 1200 U T C using the C A N E R M model. 500 mb geopotential chart for 20 January 1997 at 0600 U T C . Surface isobaric chart for 18 January 1997 at 1800 U T C . Surface isobaric chart for 25 January 1997 at 1800 U T C . 7  th  Figure 20 -  7  th  Figure 21 -  7  th  Figure 22 -  7  3  Figure Figure Figure Figure  23 24 25 26  -  Figure Figure Figure Figure  27 28 29 30  -  Figure 31 Figure 32 Figure 33 -  vii  Figure 34 Figure 35 Figure 36 Figure Figure Figure Figure Figure Figure  37 38 39 40 41 42  -  Figure 43 Figure 44 Figure 45 Figure 46 -  Figure 47 -  Figure 48 -  Three dimensional back trajectories arriving over the L F V area on 23 January 1997 at 0000 U T C using the C A N E R M model. Three dimensional back trajectories arriving over the L F V area on 23 January 1997 at 1800 U T C using the C A N E R M model. Three dimensional back trajectories arriving over the L F V area on 26 January 1997 at 1200 U T C using the C A N E R M model. 500 mb geopotential chart for 1 January 2001 at 0000 U T C . 250 mb geopotential chart for 2 January 2001 at 1200 U T C . 500 mb geopotential chart for 7 January 2001 at 1200 U T C . Surface isobaric chart for 1 January 2001 at 0000 U T C . Surface isobaric chart for 6 January 2001 at 1200 U T C . Three dimensional back trajectories arriving over the L F V area on 6 January 2001 at 1800 U T C using the C A N E R M model. Three dimensional back trajectories arriving over the L F V area on 7 January 2001 at 1800 U T C using the C A N E R M model. G O E S water vapor satellite image for 1 January 2001 at 0000 U T C . Possible stratospheric and tropospheric source regions of ozone that arrived in the L F V during the four case studies in Section 4.3. Absolute latitudinal, longitudinal and height differences between the C A N E R M and H Y S P L I T models for back trajectory runs starting at 100 m a.g.l, 500 m a.g.l. and 1000 m a.g.l. over the L F V . Daily B e activity concentrations measured at the C T B T O I M S station C A 0 0 2 and total daily precipitation measured at Vancouver International Airport during October 1996. Daily 7Be activity concentrations measured at the C T B T O I M S station C A 0 0 2 and total daily precipitation measured at Vancouver International Airport during August 1997. 7  viii  Acknowledgements I would like to thank all o f those who provided me with funding for the current research. Firstly, I would like to thank my supervisor, Dr. D o u w Steyn, who provided me with funds from the National Science and Engineering Research C o u n c i l ( N S E R C ) and the Canadian Foundation for Climate and Atmospheric Sciences ( C F C A S ) . I would also like to thank the Radiation Protection Bureau ( R P B ) at Health Canada for B e data and 7  additional funding. I am also grateful to K e n Reid at the A i r Quality Department o f the Greater Vancouver Regional District ( G V R D ) for the ozone data and the Environment Canada ( E C ) Meteorological Service o f Canada's Pacific Y u k o n Region and Canadian Meteorologial Center offices for supplying me with temperature and precipitation data for Abbotsford and Vancouver International Airports and the meteorological back trajectory data. During the course o f my research I had the opportunity to share research ideas and receive advice from a number o f people. They are, i n no particular order: Dr. Kurt Ungar, Dr. Bliss Tracy and Sonia Johnson at R P B , Dr. D o u w Steyn, Dr. David Measday and Dr. Alberto Martilli at U B C , Rene Servranckx and M i c h e l Jean at the E C Canadian Meteorological Center in Dorval, P Q . During the darker climbs, I would like to thank my parents, my girlfriend, Sara and my housemates, who tolerated me and offered me kind words o f encouragment.  ix  .1. Introduction Ozone plays an important dual role in the Earth's atmosphere. In the stratosphere, the ozone layer protects animal and plant life on the Earth's surface by absorbing shortwave electromagnetic radiation in the 0.2-0.4 urn wavelength range. Without the shielding effect of the stratospheric ozone layer, skin cancer, amongst other diseases, would increase. In the troposphere however, ozone has a much more sinister effect. Produced through a variety of precursor chemical species, both anthropogenic and natural in origin, tropospheric ozone, in sufficiently high concentrations, can cause lung damage, coughing, nausea, pulmonary congestion and can act as an eye irritant (NRC, 1992).  Tropospheric ozone can also be harmful to plant life and crops by  destroying a plant's leaves, hindering photosynthesis and slowly "starving" the plant until it eventually dies (NRC, 1992). Ozone in the troposphere has several sources, both natural and anthropogenic. The main precursor pollutants that eventually react in the presence of shortwave solar radiation and high temperatures to form tropospheric ozone are volatile organic chemicals (VOC) and oxides of nitrogen, especially N O and NO2, denoted as N O . x  The  largest  sources  of these precursor pollutants  in large  urban areas are  anthropogenic and include industrial and commercial facilities, motor vehicles and power plants.  Some lesser natural sources include lightning, biomass burning and  soil. Tropospheric ozone can also originate from the stratosphere. During intense lowpressure synoptic systems and thunderstorms, ozone from the stratosphere can pass through the tropopause into the troposphere (Johnson and Viezee, 1981). This ozone can then be transported horizontally and vertically to the surface, thereby adding to the anthropogenic ground-level ozone concentration.  1  Recently, studies have shown that sources of tropospheric ozone are larger than its sinks  (NRC, 1992).  Janach (1989) showed that the  average  annual ozone  concentration over Europe has increased by 1-2% over the past 30 years. Most of this increase can be attributed to photochemically produced ozone. The various sinks for tropospheric ozone include irreversible transport into the stratosphere, dry deposition to vegetation, soils and oceans, scavenging by N O and wet transformation in clouds x  (NRC, 1992). Elevated tropospheric ozone concentrations, especially downwind from highly populated urban areas were identified as a human health and crop degradation problem as early as the 1940's when vegetable crops showed signs of damage in the Los Angeles basin.  In later studies published in the 1950's, photochemically  produced ozone was identified as the main contributor to crop damage within the basin (Haagen-Smit et al,  1951, 1953, Haagen-Smit 1952, Haagen-Smit and Fox,  1954, 1955, 1956). Over the next 30 years, many tropospheric ozone studies were conducted worldwide (Steyn et al, 1997, Gusten et al, 1988, Juarez et al, 1994, Pont and Fontan, 2000, Rao et al, 1991). Some of the major subjects included the reaction mechanisms of photochemical pollution (Mayer, 1999, Seinfeld, 1986, Stewart et al, 1977) and air pollution modelling (Tulet et al, 2000, Morison et al, 2002, Zannetti, 1990). Although most of the tropospheric ozone originated from anthropogenic precursor pollutants (Steyn et al., 1997), there has been a recent interest in the component of ozone that was derived from natural sources, most importantly from the stratosphere (Danielsen, 1968, Davies and Schuepbach, 1994, Stohl et al, 2000). Ozone can be injected into the troposphere via stratosphere-troposphere exchange (Johnson and Viezee, 1981, Cho et al, 1999). The main mechanism of this exchange is tropopause  2  folding which involves an isentropic down folding of the tropopause which occurs behind upper or lower frontal surfaces (Danielsen, 1968).  In order for irreversible  exchange of ozone-rich stratospheric air within the fold to occur, turbulent mixing, usually in the form of Kelvin-Helmholtz waves, must occur on the lateral or bottom sides of the fold (Cho et al, 1999). Once in the troposphere, stratospheric ozone can be vertically transported to the surface either through large-scale convection (i.e thunderstorms) (Tremblay and Servranckx, 1993) or subsidence (i.e. synoptic high pressure systems) (Kato et al,  1990).  Alternately, if the bottom side of the fold  penetrates deep into the troposphere so that it impinges on the surface, ozone can be directly transferred to the ground.  This particular process will increase the  concentration of ground level ozone dramatically, but this direct transfer process usually occurs in high altitude locations only (i.e. mountainous regions) (Schuepbach et al, 1999, Stohl et al, 2000, Bonasoni et al, 1999). In order for researchers to determine the magnitude of the stratospheric ozone contribution to ground level ozone concentrations, some sort of stratospheric tracer, measurable at the surface, must be utilized. Over the years, many tracers have been used including relative humidity (Stohl et al. 2000), S r from American stratospheric 90  nuclear tests (Danielsen 1968, Danielsen et al. 1970, 1977), and potential temperature (Haagenson, 1981). Ozone itself cannot be used as a tracer because it is not produced exclusively in the stratosphere.  High surface ozone concentrations could be  anthropogenic and/or natural in origin even in remote locations. It is the previously mentioned tracers that reveal the stratospheric source.  In recent years, the  cosmogenic radionuclide, Be, has emerged as the primary stratospheric tracer in 7  stratosphere-troposphere exchange (STE) studies (Murao et al, 1990, Tremblay and Servranckx, 1993, Stohl et al, 2000, Stohl and Trickl, 1999, Bonasoni et al, 1999).  3  Its usefulness as a tracer of S T E stems from the fact that it is produced in the upper troposphere and lower stratosphere where tropopause folding occurs.  Further  characteristics of B e will be covered more thoroughly in Section 2.1.1. 7  The bulk of studies involving S T E have been conducted either in mountainous regions around the world, most importantly in the Swiss Alps, or in-situ via aircraft measurements because air of stratospheric origin has had little chance to dilute at these altitudes.  The following two sections summarize several important papers,  published over the past 30 years, which greatly aided the research community in this field. 1.1 In Situ Aircraft  Studies  Danielson et al. (1970) were among the first researchers to study the distribution of radioactivity and ozone associated with tropopause folding. They were also the first researchers to show that there was a positive relationship between radioactivity and ozone originating from the stratosphere. Their measurements were carried out by fitting four RB-57 aircraft of the U S Air Weather Service (now known as the National Weather Service) with radioactivity and ozone sensors and flying througha tropopaus foldbetween Albuquerque, N M and the southwest corner of Wyoming. It was reported that favourable meteorological conditions in which to observe a tropopause fold were an intensifying upper-level trough over western North America, a ridge over the Gulf of Alaska and a strong upper level jet that is advancing around the ridge. They also concluded that velocity shear produced by gravity waves plays an important, if not primary role in the mixing of stratospheric air into the troposphere. Johnson and Viezee (1981) investigated the frequency and extent of stratospheric intrusions in the Central U.S.A. during the spring and fall of 1978 using in-situ aircraft measurements.  At least during measurement flights, stratospheric ozone  4  intrusions were observed in nearly every trough regardless of its intensity. maximum ozone concentrations  in the intrusions and the trough intensity  The as  characterized by the maximum wind speed at 300 mb were found to be highly correlated. During these intrusion events, peak ozone concentrations, between 240400 ppb, were observed between 6-8 km a.s.l.  The ozone concentrations then  decreased to 100-200 ppb at lower levels as a result of turbulent mixing processes causing dilution. Measured ozone concentrations during spring were almost twice as high as those measured in the fall. These intrusions were generally 100-300 km wide in the crosswind direction, several hundred kilometres long and could penetrate as deep as the top of the boundary layer. They postulated that possible mechanisms for downward transport within the boundary layer include normal convective mixing, organized convection associated with cloud and precipitation processes and organized downward motion within frontal zones. 1.2 High Altitude Studies Stohl et al. (2000) analyzed climatological data of B e and ozone to characterize 7  the frequency of stratospheric intrusions at three high altitude stations in the Alps and Apennines.  These stations were Jungfraujoch, Zugspitze and Sonnblick. The data  sets indicate that the absolute frequency of stratospheric intrusions depended on the threshold values that one sets for Be, ozone and humidity. However, they noted that 7  the relative shape of the annual cycle of stratospheric intrusions is fairly insensitive to the variations in the threshold values.  Sonnblick and Zugspitze showed the highest  stratospheric intrusion frequency in October with a secondary maximum in January and February and a deep minimum in summer.  No clear seasonal stratospheric  intrusion patten could be found in the Jungfraujoch data. Zanis et al. (1999) investigated some of the factors affecting B e at Jungfraujoch 7  5  in the Swiss Alps. They used a data set spanning 4 April 1996 - 1 January 1997. It was found that downward transport associated with an upper ridge and a high tropopause is a significant controlling mechanism for high Be activity concentrations (Bqm" ) at Jungfraujoch. However, they did state that this direct process is not the 3  only way that stratospheric B e and ozone were introduced to Jungfraujoch. A n 7  indirect multi-step mechanism might also be responsible.  Zanis et al (1999) outline  such a case that occurred from 16-23 July 1996 where it was concluded that stratospheric ozone that reached the station had originated from an S T E event over northern Europe. The air that originated from this event was advected southward to Switzerland and eventually subsided to the surface. These studies outlined not only the shape and frequency of stratospheric intrusions, but also the favourable meteorological conditions in which these intrusions occur. These studies also describe the vertical and horizontal transport mechanisms required for stratospherically intruded ozone to reach the surface. 1.3 Sea Level Studies Although less prevalent in the literature, sea level case studies have also been conducted. Although the stratospheric signature of the intruded air is more dilute at altitudes close to sea level, it can be measured especially during the winter months when anthropogenically derived ozone production is at a minimum. B e is a useful 7  measurement tool in these studies, because the other stratospheric tracers have usually become too dilute to identify the air mass as partly stratospheric in origin. Lamb (1977) investigated one case in Santa Rosa, C A , on 19 November 1977 in which that city's air quality monitoring station recorded five consecutive hours of ozone concentrations above the National Ambient A i r Quality Standard of 120 ppb. During this time, the highest hourly average was 230 ppb. Lamb (1977) gives three  6  possible reasons for such high values. The first was instrument malfunction, but after some testing, this hypothesis was ruled out. The second was that these high oxidant concentrations originated from anthropogenic sources.  The fact that this event  occurred before sunrise rules out the possibility of photochemical processes unless the ozone was transported to the region.  After model trajectory analysis, it was  concluded that anthropogenic ozone was not responsible for this episode.  The third  possibility was that stratospheric ozone had made its way to the surface via S T E . Through isentropic and isothermal analysis, Lamb (1977) was able to compute the vertical velocities and determined that there was descent along a frontal zone that passed over the area. Further evidence, including potential vorticity and mixing ratio profiles, suggested that this air, and thus the ozone, was indeed of stratospheric origin. Chung and Dann (1985) document a rare high ozone concentration episode that occurred on 26-28 December 1980 at Regina, Canada in which one hour average oone concentrations ranged from 138-228 ppb. The meteorological conditions at this site were similar to other cases in which stratospheric intrusions occurred. They were :  •  The presence of a quasi-stationary front with the formation of a lee cyclone east of the Rocky Mountains.  •  The approach of an upper-level trough.  •  The passing of a warm front over the area after a cold spell followed by a rapid temperature drop associated with the passing of a cold front. Wind shifts (presumably caused by the passing of the warm and cold fronts).  • •  The presence of a strong jet-stream maximum at all levels from 850 mb to 250 mb over southern Saskatchewan.  The occurrence of such an episode  is rare for the region even though the  meteorological conditions are frequently favourable for such stratospheric intrusion events. The study conducted by Tremblay and Servranckx (1993) was one of the few studies, conducted near sea level, that measured Be and ozone simultaneously. The  7  purpose of the study was to determine what fraction of the total ozone measured in the Montreal, Canada area originated from the stratosphere. 7  On 27-28 June 1991, high  B e activity concentrations were recorded which suggested that some of the ozone at  the surface was of stratospheric origin. During these two days, two thunderstorms with cloud tops reaching the lower stratosphere hit the southwestern portion of Quebec. Davies and Schuepbach (1994) suggested that downdrafts in thunderstorms could be a possible mechanism for the transport of ozone from the stratosphere to the surface.  Tremblay and Servranckx (1993) reported that the average annual B e 7  activity in Montreal is about 3400 uBqm" . 3  The values of B e on 27-28 July 1991  were 5600 pBqm" and 6500 uBqm" respectively. 3  3  7  This suggested that some of the  ozone measured at the Montreal station, located at the top of the Atmospheric Environment Service building, originated from the stratosphere. They concluded that this stratospheric ozone could have acounted for 18% (or 9 ppb by volume) of the total ozone measured at the surface during the measurement period. 1.4 Research Goals Although several studies have used B e data to investigate single, high ozone 7  cases (Murao et al,  1990, Tremblay and Servranckx, 1993, Bonasoni et al, 2000),  very little attention has been paid to an examination of the climatology of both species ( Be and O 3 ) simultaneously. This was mostly due to non-existent or at best sporadic 7  7  B e measurements.  Since the implementation of the Comprehensive Nuclear Test  Ban Treaty (CTBT) (see Section 2.1.1), B e measurements are now readily available 7  in many locations worldwide.  The current research will make use of B e data 7  collected at a C T B T radionuclide monitoring station so that, in conjunction with a regional ozone monitoring network, cases in which stratospheric ozone is believed to  8  have had an influence on ground level ozone concentrations, may be identified and analysed. The current research utilizes data collected in the Lower Fraser Valley (LFV) in southwestern British Columbia, Canada.  It has long been known that this region  suffers from a moderately acute ozone pollution problem (Steyn et al., 1997). Although most of the ozone originates from anthropogenic precursor emissions, a lesser, but certainly not negligible  contribution may b originating from the  stratosphere. This stratospheric contribution will be the focus of the current research. The goals of the current research are: •  To identify episodes, utilizing extensive B e and ozone data sets, in which 7  stratospheric ozone, via S T E , contributes to the ground level ozone concentrations in the L F V . •  To identify source regions of stratospheric ozone which influences ground level ozone concentrations in the L F V .  Two Lagrangian trajectory models, the  Canadian Emergency Response Model ( C A N E R M ) and the N O A A A i r Research Laboratories Hybrid Single Particle Lagrangian Integrated Trajectory model (HYSPLIT) will be utilized to determine not only source regions, but also the transport pathways of this stratospherically intruded ozone from the time of intrusion to the time of arrival in the L F V . •  To select four episodes from the identified stratospheric episodes and perform a detailed analysis to determine the various synoptic conditions necessary not only for S T E , but also for the vertical and horizontal transport to the L F V .  •  To qualitatively determine how precipitation affects B e concentrations in the 7  LFV.  9  The Be, ozone and temperature data sets for the current research span the period 7  10 April 1996 to 12 March 2002. The CTBT radionuclide station at which Be is 7  measured is located at the University of British Columbia (Figure 11). Ozone concentrations are measured at several stations which are part of the Greater Vancouver Regional District air quality monitoring network (Figure 11). Other data such as 300 mb and 500 mb geopotential height charts, surface pressure charts, precipitation and infrared and water vapour satellite images will be utilized where possible during the case studies as they are not available continuously throughout the study period. 2. Data 2.1 Be in the Atmosphere 7  Early stratosphere-troposphere exchange research utilized fissionogenic radionuclides as stratospheric tracers, most notably Sr and 90  137  Cs (Danielsen, 1968).  These  radionuclides were products of nuclear tests that were conducted both in the atmosphere and on the ground.  These radionuclides were injected into the  stratosphere during a nuclear blast in the case of ground tests or were generated in situ in the case of stratospheric tests.  In time, Sr and 90  137  Cs concentrations in the  atmosphere and in the stratosphere in particular began to decrease, either because of dry deposition to the earth's surface, scavenging due to precipitation, half life decay or ceased nuclear testing (Ishikawa et al, 1995). Researchers eventually settled on the cosmogenic radionuclide Be (half life of 53.3 days) due to its abundance in the 7  lower stratosphere and its steep concentration gradients between the stratosphere and the troposphere (Lai and Peters, 1967). 7  12  Be is formed by spallation processes of light atmospheric nuclei such as C, 1 4  N and  16  0 with protons and neutrons (Zanis etal, 1999). The galactic particles 10  BCP 60° North  40°  20*  0  9  20°  40°  «J*  SC  South  Geomagnetic Latitude  Fi gurc 1 - A modeled north-south section through the atmosphere showing the surfaces of constant Be production. Units are in nuclei min" m" at standard temperature and pressure. Taken from Lai and Peters (1967). 7  1  3  approach the solar system isotropically and tend to be deflected by the Earth's magnetic field and travel along field lines; hence, particles lose energy after successive collisions, and can (with low energy) enter the atmosphere predominantly over the magnetic poles, resulting in a pronounced pole-equator gradient in the flux of less energetic particles into the atmosphere which similarly leads to a pole-equator 7  7  gradient of Be production (Zanis et al, 1999). About 75% of Be is produced in the stratosphere and 25% in the troposphere (Johnson and Viezee, 1981) and there is a strong concentration gradient between the stratosphere and the troposphere (Feely et al, 1989). Figure 1 shows the modelled Be production rates in the atmosphere. 7  It  can clearly be seen from the figure that the production rates in the polar stratosphere are very large compared to the rest of the atmosphere and that steep horizontal and vertical gradients exist away from these locations.  Talpos and Cuculeanu (1997)  modeled vertical concentration profiles of B e at 45°N. From the graph in Figure 2, it 7  can be seen that B e has a maximum concentration at approximately 15 km which 7  11  z{m)  c{10*nuclm ) J  Figure 2 - The modeled vertical concentration profiles of Be at 45°N for settling velocities (1) u =0 ms" , (2) u = 0.02 ms" and (3) u = 1 ms' . Taken from Talpos and Cucleanu(1997). 7  1  1  s  s  1  s  then rapidly decreases to nearly zero at the surface. Seasonal variation of surface Be activity concentration varies greatly from place 7  to place. Caillet at al. (2001) identified a summer maximum in Geneva, Switzerland. Bovis and Steyn (2001) and Megumi et al. (2000) determined a fall and spring maximum in Vancouver, Canada and Osaka, Japan respectively. Paatero and Hatakka (2000), Hotzl and Winkler (1987) and Dibb et al. (1994) identified spring maxima in Sodankyla, Finland; Neuherberg, Germany and Alert, Canada respectively.  This  7  difference in the seasonal variation of  7  Be is due to several factors.  Since  Be  concentrations at the surface greatly depend on air masses that have been transported from the upper troposphere/lower stratosphere, surface B e concentrations are closely 7  linked to the seasonality of STE (Feely et al, 1989). Another major factor is that B e 7  attaches to sub-micron sized particles shortly after formation (Ishikawa et al, 1995), which makes B e susceptible to scavenging by precipitation. 7  Feely et al. (1989)  found strong correlations between average monthly precipitation and surface B e 7  concentrations  for stations that had strong seasonal variations in precipitation.  12  180" -150" -120' -90' -60" -30" 0'  30' 60' 90' 120" 150' 180'  180" -150' -120" -90' -60' -30' 0  30' 60' 90" 120' 150' 180'  -  Figure 3 - Map of the Comprehensive Nuclear Test Ban Treaty Internaional Montoring System radionuclide network. Taken from Medici'(2001). Ishikawa et al. (1995) showed that only a small amount of precipitation was required to scavenge a sizeable amount of B e out of the atmosphere. 7  deposition measurements  Thus, if only dry  are being taken, a significant amount of information  regarding B e surface concentrations is being lost. 7  2.1.1 Radionuclide Measurements as part of the CTBT The Comprehensive Nuclear Test Ban Treaty (CTBT) was a milestone agreement signed by 71 countries on 10 September 1996.  The C T B T stated that no nuclear  detonation of any kind or in any environment (air, land, underwater or underground) was to take place at any time. To ensure the treaty's compliance, the Comprehensive Nuclear Test Ban Treaty Organization (CTBTO) set up an International Monitoring System (IMS) network of 321 hydro acoustic, seismic, infrasonic and radionuclide monitoring stations and 16 radionuclide laboratories across the globe (Figure 3) to monitor and trace the origin of any nuclear detonations (Medici, 2001). The radionuclide network is composed of 80 stations containing high volume air samplers, which collect aerosols and particulates of sub-micron size or larger on glass  13  fiber filters (Measday et al., 2001). These filters are then analyzed for radionuclides, especially those of fissionogenic origin. A side benefit of the radionuclide network is that many naturally occurring and anthropogenically  produced radionuclides  are monitored by  the  IMS as well.  Bovis and Steyn (2001) and Measday et al. (2001) have already utilized these radionuclide  data  in  the  fields  of  stratosphere  troposphere  exchange  and  anthropogenic radionuclide monitoring respectively. 2.1.2 Be Measurements in Vancouver, Canada 1  The IMS radionuclide station CA002 is located at The University of British Columbia west of Vancouver (49.25° N , 123.25° W). The station is located in a selfcontained trailer next to fields used by the Faculty of Agricultural Sciences on a plateau about 50 m above sea level. The data are collected using a high volume air sampler capable of 1000 m hr"'. Air is drawn through a high efficiency particulate 3  filter with a square face area of 60 cm x 60 cm. Every 24 hours, the filter is removed and compacted into a disk with a diameter of 6.4 cm and a thickness of 1.4 cm (Measday et al, 2001). The disk is then placed between two high-purity germanium detectors.  The data collected by the germanium detector are then sent to the  International  Data  Center  in  Vienna,  Austria  for  further  analysis  (http://www.ctbto.org). The B e data utilized for this study range from 9 April 1996- 12 March 2002 and 7  have a temporal resolution of 24 hours. The entire Be time series is shown in Figure 4.  In agreement with Hotzl and Winkler (1987), the data are closely log- normally  distributed (Figure 5). The geometric mean and standard deviation are 2032 uBqm" and  1773  uBqm"  3  respectively.  Figure 6 shows the  monthly  average  7  Be  concentrations recorded at this station. The composite seasonal time series (Figure 6)  14  o  —i J  1 S  i D  1 M  1 J  1996  r—i S  1997  D  1 M  1 J  1 S  1  D  M  1998  1  1 J  S  1 D  r—i M  J  1999  1  1  1  1  1  S  D  M  J  S  2000  1 D  2001  Figure 4 - Daily B e activity concentrations measured at the C T B T O I M S station C A 0 0 2 from 9 April 1996 - 12 March 2002. 7  loooog B 7  Cumulative D i s t r i b u t i o n  (%)  Figure 5 - Cumulative distribution of the B e activity concentrations at the C T B T O IMS station CA002 during the period 9 April 1996- 12 March 2002. 7  15  1 1 M  Figure 6 - Monthly average B e concentration at the C T B T O IMS station C A 0 0 2 during the period 9 April 1996- 12 March 2002. 7  Figure 7 - Average number of days in a month in which B e activity concentration exceeds the 80th percentile at the C T B T O IMS station C A 0 0 2 during the period 9 April 1996 - 12 March 2002. 7  16  Change in Odd Oxygen 0  2  + /ro-+0 + 0  0 +0 0  3  +  2  + M->0  +2 3  ho -> O + 0  O +0  3  - * 20  +M  0 0  2  -2  2  0 +0 +M ^ 0  2  +M  -2  Figure 8 - Chemical reactions of Chapman's mechanism, a static pure oxygen photochemical steady state model.  is bimodal with a spring and fall maximum and a summer and winter minimum. It must be kept in mind that these data are based on dry deposition measurements only so wet deposition caused by scavenging of B e by precipitation is not taken into 7  account. The winter season in Vancouver is associated with high precipitation, so if wet deposition measurements were being taken, the winter B e concentration would 7  be significantly higher. Figure 7 shows the average number of days in each month in which B e exceeds the 80 7  th  percentile. Again, spring and fall maxima are observed  along with a secondary maximum in winter. These are periods in which air masses associated with stratosphere troposphere exchange could have an influence on Be measurements taken at this station.  2.2 Ozone 2.2.1 Ozone Formation in the Stratosphere Chapman (1930) was the first scientist to produce a theory of ozone formation in the stratosphere. Chapman's approach used oxygen as the only element involved in the reaction. Figure 8 shows the five reactions proposed by Chapman, along with the change in odd oxygen for each reaction, of the formation and destruction of ozone in the stratosphere.  It was later found that this pure-oxygen reaction model produced  ozone concentrations that were higher than observational values (Prinn et al, 1978).  17  X +  o -> xo + 0 3  X O + O -> X + 0  o+0  3  -> 0  2  2  2  + 0  2  Figure 9 - A generic sequence of reactions for the loss of odd oxygen by way of a catalytic cycle. 60  Altitude (km)  50  -- — Measured  •  - — OknUlnl  .KPN  40 30 20  ^  (ffiN  \  ^ \  IK**"* •  10 ******  0  « M . ^ I ^ U . ••ii f n T pf__.  10" 10"  10''  10"  10  1J  10"  IO'  10  13  3  10  10  1?  10"  1  (45°N)  00°N)  Concentration (molecule cm-3) Figure 10 - Observed (solid line) and calculated (dashed line) vertical concentration profiles of ozone as a function of latitude in the Northern Hemisphere for midspring. Taken from Miller et al. (1981). It was also found that the fourth reaction in Figure 8 only accounted for a quarter of the actual ozone loss (Wayne, 1999). Bates and Nicolet (1950) proposed that trace atmospheric constituents could participate in catalytic process that removes ozone at a faster rate than the pureoxygen model put forth by Chapman. The essence of catalytic schemes for the loss of odd oxygen is the provision of a more efficient route than the last two reactions in Figure 8 (Wayne, 1999).  The chain mechanism depicted in Figure 9 achieves the  same result as the fourth equation in Figure 8, but because of the catalyst, the reaction becomes faster, and ozone is destroyed more quickly than the Chapman mechanism (Wayne, 1999). Various chemical species have been identified as the catalyst, X , in  18  Figure 9 including H , O H , N O and C l (Wayne, 1999). Miller et al. (1981), using a complex model based on catalytic cycles, found good agreement when compared with measured ozone profiles (Figure 10).  2.2.2 Ozone Formation in the Troposphere In an urban region, ozone is formed in two major ways. The largest is the result of a series of reactions involving anthropogenically produced N O  x  and volatile  organic chemicals (VOCs) in the presence of ultraviolet radiation from the sun (Seinfeld and Pandis, 1998). Since temperature is a controlling factor on the reaction rate, the higher the temperature, the higher the measured ozone concentration. Anthropogenic ozone production is thus at a maximum during summer.  A less  important mechanism is a series of reactions involving naturally produced N O and x  V O C s in the presence of ultraviolet radiation (NRC, 1992). Wayne (1999) and N R C (1992) give a more detailed discussion on ozone formation in the troposphere with special attention paid to photochemical smog of which ozone is a major constituent.  2.2.3 The Greater Vancouver Regional District Air Quality Monitoring Network Because of the geography of the Greater Vancouver Regional District, the region suffers from an acute ozone pollution problem, especially in the summer (Steyn et al., 1997).  To the west of the G V R D lies Georgia Straight which is a salt water body  separating the British Columbia mainland and Vancouver Island. Mountain ranges lie to the north and south and the valley between these two ranges, in which the G V R D is located, becomes narrower further east.  As pollutants travel down valley, they  become trapped between the two mountain ranges as the valley narrows (Steyn et al., 1997).  Because of this, air pollutant concentrations, especially ozone, have a  tendency to increase eastward in the G V R D .  19  Figure 3 - Greater Vancouver Regional District Continuous Air Quality Monitoring Network. Air quality monitoring stations are indicated by the circles. Abbotsford International Airport is located at station T33, Vancouver International Airport is located at station T31 and the CTBTO IMS station CA002 is located just to the east of station T2. The stations used in this study are: Downtown Vancouver (Tl), Kitsilano (T2), Chilliwack Airport (T12), Surrey East (T15), Richmond (T17), Langley Central (T27), Hope Airport (T29) and Port Coquitlam North (T32).  In order to observe the air quality in the G V R D , a monitoring network was set up starting in 1949. Today, there are over 30 monitoring stations operated by the G V R D Air  Quality  Department,  most  of which  measure  hourly concentrations  of  V O C s , N O , ozone, C O , SO2 and suspended particulates. This monitoring network is x  displayed in Figure 11. The G V R D ozone data set associated with this research spans between 1 January 1996 - 31 December 2001. maxima and a winter minimum.  A l l of the stations exhibit spring-summer  Figure 12 contains composite diurnal ozone  concentration graphs for all stations used in this study. These stations are Downtown Vancouver (T001), Kitsilano (T002), Chilliwack Airport (TO 12), Surrey East (TO 15), Richmond (TO 17), Langley Central (T027), Hope Airport (T029) and Port Coquitlam North (T032).  The error bars denote ± one standard deviation. These composites  20  a).  21  c).  22  23  24  Figure 12 (previous pages) - Composite diurnal ozone concentration graphs for a). Downtown Vancouver (T001), b). Kitsilano (T002), c). Chilliwack Airport (T012), d). Surrey East (T015), e). Richmond (T017), f). Langley Central (T027), g). Hope Airport (T029) and h). Port Coquitlam North (T032). The error bars denote +/- one standard deviation. The composites were constructed by averaging all of the measurements for a given hour at a given station.  were constructed by averaging all of the measurements for a given hour at a given station. The resulting graph is an average diurnal ozone time series for each station, each of which exhibits a daily maximum at 2:00pm PST and a daily minimum between 7:00am - 8:00am PST.  2.3 Temperature and Precipitation This research makes use of hourly measurements of temperature and precipitation taken by the Meteorological Service of Canada spanning 1 January 1996 - 31 December 2001.  The temperature  data set includes data from Vancouver  International Airport and Abbotsford International Airport while the precipitation data set includes data from Vancouver International Airport only. These two stations are labeled in Figure 11.  2.4 Satellite Data Several satellite images were extracted from the Canadian Meteorological Center database for use with specific case studies covered in Section 3 of this thesis. The thermal  infrared data and water vapor data were measured by the N A S A  Geostationary Earth Orbiting Satellite (GOES-10) with a pixel resolution of 4km. The thermal infrared data were measured at 10.7 microns and the water vapor data were measured at 6.7 microns. The temporal resolution of this data set is three hours.  3. Data Analysis There are two main goals of the current research. The first is to identify episodes where ozone transported from the stratosphere is possibly influencing ground level ozone concentrations in the L F V . The second goal is to utilize meteorological back  25  trajectory analysis to identify the source regions of this stratospheric ozone.  Once  these episodes have been identified, several case studies will be investigated in greater detail and will include, in addition to B e and ozone analysis, an examination of the 7  synoptic meteorology as well as temperature and precipitation measurements. This section will describe the two algorithms that were used to detect episodes in which air of stratospheric origin may be influencing the ground level concentrations in the L F V .  ozone  Section 4 introduces the concept of meteorological back  trajectory analysis and the models that were used in the current research. Section 4 will also cover the four detailed case studies selected from the episodes identified in Sections 3.1 and 3.2.  3.1 80th Percentile Algorithm This algorithm selected the 80 and the  80  th  percentile  th  percentile of all daily B e activity concentrations 7  of all average  daily maximum ground level  ozone  concentrations. The 80 percentile cutoff was selected because at cutoff percentiles th  lower than this,  some of the resulting cases may not contain  7  Be  activity  concentrations and ground level ozone concentrations that are being influenced by air of stratospheric origin. By picking cutoff percentiles higher than 80% some possible cases of stratospheric intrusions may be missed. On a given day, the average daily maximum ozone concentration is found by selecting and then averaging the highest hourly ozone concentration from each station across the network. The intersection of these two 80 percentile data sets was then found so that the final resulting data set th  contained days in which both the daily B e activity concentration and the average 7  daily maximum ground level ozone concentration were in the 80  th  percentile.  This  algorithm was then modified twice to exclude months in which anthropogenically derived ozone production is at a maximum.  26  These modified algorithms selected  episodes in which stratospheric ozone may be influencing ground level ozone concentrations in the Lower Fraser Valley.  3.1.1 80th Percentile Algorithm Using All Data A l l of the cases that were detected using this algorithm were distributed among 54 episodes, each episode possibly containing more than one case. A case is a period in which both the B e activity concentrations and the average daily maximum ozone 7  concentrations are in the 80 percentile during a single day or on consecutive days. If th  more than one case occurred over the span of a few days,  1 2 4 5  Max Be (pBqm- )  Max [0 ] (PPb)  3924.5 5750.2  40.0 48.2  7  Episode  Dates (yy/mm/dd) 96/4/29, 96/5/1 96/5/8-96/5/9, 96/5/11  3  3  6614.8 5613.9  49.4  6  96/5/15-96/5/17 96/6/2/-96/6/3 96/6/15, 96/6/21  4234.5  40.8  7  96/7/2, 96/7/8, 96/7/14  5332.4  48.7  8  96/7/26, 96/7/28-96/7/29  6677.1  68.5  9 10  96/8/9-96/8/10 97/3/30  5259.4 4961.8  69.2 42.9  11  97/4/6-97/4/8, 97/4/11  5686.4  44.3  12  97/4/20, 97/4/26  4208.2  44.8  13  9290.6  67.3  14  97/5/3, 97/5/9-97/5/16, 97/5/18-97/5/19 97/6/6  4638.3  40.6  15  97/7/4  5887.8  45.2  16  97/7/18-97/7/20  5677.3  61.1  17  97/8/10-97/8/12  6835.3  49.8  18  98/4/9, 98/4/13, 98/4/17-98/4/18  5496.2  42.7  19  6381.5  46.1  20  98/4/20-98/4/21, 98/4/28-98/4/30 98/5/1-98/5/4  7582.0  51.1  21  98/7/29, 98/8/4  4376.0  69.5  22  98/9/5-98/9/7  8077.3  42.2  23  99/3/1  4316.1  37.3  24  99/4/2  4036.8  42.7  41.6  25  99/4/15-99/4/18  6309.8  50.9  26 27  99/4/30 99/5/6  4796.0 6070.8  42.1  28  99/5/24, 99/5/27  6211.0  54.9  29  99/6/12-99/6/14  5867.5  43.9  30  99/7/9-99/7/11  5771.7  58.2  27  39.1  31  99/8/23  3668.2  37.7  32  99/9/29  3680.1  48.9  33  00/4/9  3821.6  41.2  34  00/4/17-00/4/21  10845.5  47.2  35  00/4/27, 00/4/30-00/5/1, 00/5/3  8459.8  42.2  36  00/5/14-00/5/18  5661.2  37  00/5/24-00/5/25  5954.7  46.5 48.6  38  00/6/2-00/6/4  8827.7  53.6  39 40  00/6/17 00/6/27-00/6/28  40.5 56.2  41  00/7/30  3739.3 4960.3 3817.4  42 43  00/8/23 01/2/16  3952.5  39.5  4739.7  38.8  44  01/3/19  3716.4  39.5  45 46  01/3/25 01/4/15  405.4.3 3832.1  40.7 47.7  47  01/4/30  3967.0  50.0  48  01/5/4  5374.7  49  01/5/7-01/5/12  40.2 50.4  50  01/5/27-01/5/28  51 52 53  01/6/21 01/7/2-01/7/3, 01/7/7 01/8/10-01/8/11  4117.0 4700.6 4853.4  44.6 55.1  54  02/2/24  5418.9  38  7760.6 4938.8  .  39.4  50.4  58.0  Table 1 - Episodes based on the 80th percentile detection algorithm (all data). The highest daily Be activity concentration and the highest average daily maximum ozone concentration are indicated for each episode. 7  these cases were grouped to form one episode, thus making analysis easier. Therefore, an episode is a time period in which one or more cases occurred. Table 1 lists all of the episodes and the dates of the cases that they contain. Table 1 also lists the highest B e activity concentration and the highest average daily maximum ozone 7  concentration that occurred during each episode.  Figure 13 shows the daily B e 7  activity concentration, the average daily maximum temperature, the average daily maximum ground level ozone concentration and the coefficient of variation of ozone of four of these episodes. Graphs of all the episodes can be found in Bovis and Steyn (2001).  28  a).  29  c).  610000  < / \>  8000  3  i l 6000 > •H  0 4000 ' 1  \  H  \  / V  2000 '  13  15  17  19 April  100 I 90 i 80 ' 70 60 , 50 40 30 ' 20 10 0  GM  23  2  17  2000  <  17  0  80%ile  \  21  2  19 April  21 2000  >  19 April  21  17  2000  19 April  21 2000  d). 12000 "  30  610000 "  o 25 U  CD  5  8000 '  -H 6000 >  0 H  1  1  12  13  14  1  r—>—,  ,  15 16 17 18 February 2001  ,  1  19  20  0 -I  1  .  .  12  13  14  15  16  i  ,  ,  1  17  18  19  20  February 2001  30  Figure 13 (previous pages) - Be activity concentration, average daily maximum temperature, average daily maximum ozone concentration and coefficient of variance of ozone for a) 3 May 1996 - 22 May 1996, b) 13 July 1997 - 25 July 1997, c) 13 April 2000 - 25 April 2000 and d) 11 February 2001 - 20 February 2001 from the 80 percentile detection algorithm (all data). A single vertical line represents an episode in which both Be and ozone were above the 80th percentile for one day only. Two vertical lines with an arrow between them indicate a multi-day episode. The graph on the upper left shows the daily Be activity concentration. The top horizontal line indicates the 80th percentile value. Values above this line constitute the top 20% of all daily Be activity concentrations in the data set. The lower line (marked GM) is the geometric mean daily Be activity concentration for the entire data set. The graph on the upper right shows the average daily maximum temperature, which is found by selecting and then averaging the daily maximum temperatures measured at both Abbotsford International Airport and Vancouver International Airport (Figure 11). For the graph on the lower left, the solid line is the average daily maximum ground level ozone concentration. The upper dotted line is the highest hourly ozone concentration recorded over the entire network on a given day. The lower dotted line is the lowest daily maximum value recorded over the entire network on a given day. The upper horizontal line indicates the 80th percentile value, the middle line represents the median value and the lower line represents the 30th percentile value. The graph on the lower right shows the coefficient of variation (CV) of ozone which is the standard deviation of the daily maximum ground level ozone concentrations recorded across the network divided by the average daily maximum ground level ozone concentration. 7  th  7  7  7  7  In Figure 13, if the C V is small, the daily maximum ground level ozone concentrations recorded across the network will be similar.  Because air masses  influenced by stratosphere troposphere exchange are 100's of km long and 10's of km wide (Johnson and Viezee, 1981), one would expect a network wide rise in ozone once this air mass reached the surface. This would lower the C V value, because all monitoring stations would be recording approximately the same ozone concentration at any given time during the episode. During most of the episodes three things occur. First, a local maximum occurs in the average daily maximum ground level ozone concentration. Second, the maximum daily B e activity concentration measured during an episode corresponds to a local 7  spatial maximum in average daily maximum temperature.  This may be due to B e 7  being vertically transported to the lower troposphere by subsiding air in a high pressure system followed by entrainment into the boundary layer. This subsiding air is adiabatically warmed as it descends, thus the increase in surface temperature. Finally, the entire episode is characterized by a relative minimum in the C V of ozone. This characteristic is only evident in about 50% of the identified cases. One reason  31  for this is that the stratospheric signature of the subsiding air in some episodes may be weak. This weak signature could be due to either dilution of stratospheric air over a long transport path or a weak stratosphere troposphere exchange episode initially (or both). This weak stratospheric signature would mean that stratospheric ozone from this subsiding air could be partially or completely scavenged by N O at some stations, x  especially during the night-time hours when N O concentrations are high (NRC x  1992). This would increase the C V of ozone because the spatial structure of the daily maximum ground level ozone concentrations would increase. The disadvantage of the 80  th  percentile algorithm using all of the data is that it  selects many episodes in which ground level ozone concentrations are heavily influenced by ozone that is derived from anthropogenic precursors. A n episode in which anthropogenically derived ozone is a factor occurs when the average daily maximum temperature exceeds approximately 20°C.  Wunderli and Gehrig (1991)  showed that at temperatures higher than this, the daily maximum ozone concentration at two air quality sites surrounding Zurich, Switzerland approached or exceeded 80 ppb.  This is close to the Canadian Air Quality Objective maximum allowable one-  hour ozone concentration of 82 ppb. It was concluded that the majority of this ozone was  produced from anthropogenic  precursor pollutants  because of the  high  concentrations of peroxyacetyl nitrate (PAN) that existed during this time. During a summer S T E episode in Montreal, Tremblay and Servranckx (1993) estimated that the stratospheric contribution to the ground level ozone concentration was only about 10 ppb.  This constituted about 18% of the total ground level ozone concentration.  Therefore selecting episodes during days in which the average daily maximum temperature is lower than 20°C would ensure that stratospheric ozone would be  32  Figure 14 - Average number of days in a month in which the average daily maximum temperature exceeded 20°C in the Lower Fraser Valley during the period 10 April 1996-31 December 2001. making a larger contribution to the total ground level ozone concentration, when compared to ozone produced from anthropogenic precursor pollutants.  3.1.2 80th Percentile Algorithm Excluding June to August In order to eliminate the majority of the high temperature cases, the 80th percentile algorithm detection algorithm was refined slightly. selected the two 80  th  Before the algorithm  percentile data sets of B e and ozone, all months between June 7  and August inclusive were excluded from the data set. These months were excluded because they contained the highest frequency of days in which the average daily maximum temperature exceeded 20°C (Figure 14). 7  Be and ozone 80  th  The algorithm then selected the  percentile data sets. The intersection of these two data sets was  then found. The episodes that were selected using this algorithm are listed in Table 2.  33  Max Be (pBqin )  Max [03] (Ppb)  4129.0 7626.4  48.2  6614.8 4961.8  49.4 42.9  10718.8  47.2  4208.2  37.4  8 9 10 11 12  97/4/6-97/4/8, 97/4/11-97/4/13 97/4/26 97/5/3 97/5/9-97/5/16, 97/5/18-97/5/19 98/4/15, 98/4/17-98/4/18, 98/4/20-98/4/21 98/4/30 98/5/1-98/5/4  5764.5 9290.6 6381.5  45.5 67.3 45.5 40.2 51.1  13 14 15 16  98/9/6-98/9/7 99/2/13 99/3/8 99/3/17  17 18  99/4/2  8077.3 5732.6 4316.1 4348.3 4036.8 6309.8 4796.0 6070.8 6211.0 4089.5 4375.9 10845.5  62.8 36.2 37.3 36.8 42.7 50.9 42.1  8439.8 4799.6 5661.2  41.3 42.2 48.2 48.6  7  Episode  Dates (yy/mm/dd)  1  96/4/30  2  96/5/8-96/5/10 96/5/15-96/5/17 97/3/30  4 5 6 7  19 20 21 22 23 24  -  99/4/15,-99/4/17-99/4/18 99/4/30 99/5/6 99/5/27 99/9/14  3  .  8320.9 7582.0  36.2  39.1 49.8 37.1 36.9 47.2  25 26 27  00/4/10 00/4/17-00/4/19, 00/4/20-00/4/21 00/4/27, 00/4/30 00/5/1, 00/5/3 00/5/14-00/5/18  28  00/5/24-00/5/25  5954.7  29  4739.7  38.8  30  01/2/26 01/3/25  4054.3  40.7  31 32  01/4/30 01/5/4, 01/5/7-01/5/12  3967.0 7760.6  50.0 50.4  33  01/5/27-01/5/28  4938.8  50.4  34 35  01/10/6 02/2/24  6346.9 5418.9  37.2 38.0  Table 2 - Episodes based on the 80th percentile detection algorithm (excluding June August). The highest daily Be activity concentration and the highest average daily maximum ozone concentration are indicated for each episode. 7  Although a few high temperature cases remain, most of the episodes are low temperature.  34  3.1.3 80th Percentile Algorithm Excluding May to September The algorithm can be further refined by removing all months between May and September inclusive before starting the algorithm.  Excluding these months will  further eliminate days in which the average daily maximum temperature exceeds 20°C (Figure 14).  The resulting episodes from this algorithm are shown in Table 3.  Although the final 80  th  percentile algorithm produces only half as many cases as the  first algorithm, the number of high temperature episodes (>20°C) has been reduced from 58% to 17%. The ground level ozone concentrations associated with these low temperature episodes in the final 80  th  percentile algorithm are possibly strongly  influenced by air of stratospheric origin.  3.1.4 Disadvantages of the 80th Percentile Algorithm There is still a major problem with the third 80 percentile algorithm in that all of th  the episodes, with one exception, occur only during March and April. This is due to the fact that both ground level ozone  concentrations and daily B e activity  concentrations are quite high during these months.  7  By using the average daily  maximum ozone data instead of the hourly data recorded at the stations, one cannot obtain a detailed sense of what is happening spatially and temporally over the entire network.  The C V of ozone gives some insight on spatial variability, but does not  indicate how the concentration varies over a small temporal scale (i.e. on the order of hours).  Also, none of the 80  th  percentile algorithms describe which stations have  similar values at any one time and where the stations are located (i.e. urban or rural areas). The algorithm described next attempts to correct these three problems.  35  Max Be (uBqm- ) 7  Episode  Dates (yy/mm/dd)  1  96/4/29  2 3 4  97/3/30, 97/4/3-97/4/4, 97/4/6-97/4/7 97/4/11-97/4/13 97/4/26  5  98/4/9  6 7 8 9  98/4/13, 98/4/15, 98/4/17-98/4/18 98/4/20-98/4/22 98/4/28-98/4/30 99/2/13  10 11 12  99/3/8 99/3/17 99/4/2  13 14 15  99/4/15-99/4/18 99/4/30 00/3/8  16 17 18  00/4/10 00/4/17-00/4/21  3  M a x [03] (PPb)  3924.5 5686.4  37.5 44.3 47.2 37.4  10718.8 4208.2 3954.7 5496.2  43.7  8176.9 8320.9 5732.6  45.5 46.1 36.2  4316.1  37.3 36.8  43.3  4348.3 4036.8 6309.8 4796.0 6040.6  19 20 21 22  00/4/27, 00/4/30 01/2/16 01/3/25 01/4/15 01/4/30  4375.9 10845.5 8439.8 4739.7 4054.3 3832.1 3967.0  23 24  01/10/6 02/2/24  6346.9 5418.9  42.7 50.9 42.1 36.0 36.9 47.2 41.3 38.8 40.7 47.7 50.0 37.2 38.0  Table 3 -Episodes based on the 80th percentile detection algorithm (excluding MaySeptember). The highest daily Be activity concentration and the highest average daily maximum ozone concentration are indicated for each episode. 7  3.2 Be Only Algorithm 7  In order to alleviate the problems inherent in all three of the 80  th  percentile  algorithms, it is necessary to select the days in which the B e activity concentration is 7  high and then analyze the ozone data. For this algorithm a threshold B e activity 7  concentration of 6000 uBqm" was selected. This threshold activity concentration is 3  based on previous studies by Dutkiewicz and Husain (1979).  The 6000 uBqm"  3  threshold activity concentration algorithm identified 51 episodes, which are listed in Table 4.  Figure 15 shows the daily B e activity concentration, the average daily 7  maximum temperature, the average daily maximum ground level ozone concentration  36  Max Be (pBqm- )  M a x [03] (PPb)  96/5/9-96/5/10, 96/5/12, 96/5/15-96/5/17 96/7/3  9456.5  49.4  6007.4  5  96/7/28 96/9/11-96/9/12  6677.1 . . 7586.3  22.3 68.5 33.1  6  96/10/1-96/10/3, 96/10/7-96/10/8  11673.0  21.6  7  96/10/30-96/10/31 96/11/4  12978.6  22.2  96/12/26-96/12/27 97/1/13-97/1/16  9601.0 7022.6 8658.5  19.2 23.4  97/1/27 97/2/5-97/2/11  8623.4 12937.3  97/4/12-97/4/13  10718.8 9290.6 7451.0 6835.3  16.9 19.8 47.2  7  Episode 1 2 4  8 9 10 11 12 13 14  Dates (yy/mm/dd)  3  17.9  15 16  97/5/10-97/5/15, 97/5/19-97/5/20 97/6/12 97/8/10-97/8/11  17 18  97/9/23-97/9/25 97/10/2  19 20 21 22  97/10/19-97/10/22 97/11/14 97/12/3-97/12/5 98/3/20  6067.5 8213.1 6332.2  8.7 9.6 32.2  23 24  98/4/21-98/4/23 98/4/29-98/5/2  8176.9 8320.9  25 26 27  98/9/7 98/12/21 99/4/18  8077.3 6370.3 6309.8  45.5 51.1 40.4 28.7 43.6  28  99/5/6  6070.8  39.1  29 30  99/5/27 99/11/3  6211.0 7449.4  31 32  00/1/30 00/2/5  6617.7 6254.9  49.6 27.8 23.7  33  00/2/14-00/2/15, 00/2/17-00/2/20  8519.7  34  00/3/8-00/3/9  8267.3  36.0  35  00/4/18-00/4/20  10845.5  47.2  36  00/4/27, 00/4/30  8439.8  37  00/5/26  6814.2  41.3 33.2  38  00/6/3-00/6/5  8408.7  53.6  39  00/9/27-00/9/28  7355.9  40 41  00/10/7-00/10/8 00/12/14  8865.9 6181.9  33.1 27.2 18.5  42  01/1/3  43  01/1/8  6037.8 8980.1  20.9  44  01/1/24  6233.0  11.2  45  01/3/23-01/3/24  8366.5  32.9  8197.8 6824.6 9078.2  37  67.3 23.7 49.8 32.1 28.4 15.8  13.6 30.3  11.2  46  01/5/8, 01/5/11-01/5/12  7760.6  49.9  47  01/10/5-01/10/6  9118.6  37.2  48  01/11/10-01/11/12  7080.7  17.6  49 50  01/11/19  8334.9  02/2/16-02/2/17  8168.5  29.7 20.4  51  02/2/26-02/2/27  7553.7  31.3  Table 4 - Episodes based on the Be detection algorithm. The highest daily Be activity concentration and the highest average daily maximum ozone concentration are indicated for each episode. 7  7  and the coefficient of variation of ozone of four of these episodes. Graphs of all the episodes can be found in Bovis and Steyn (2001). Upon inspection of the graphs in Figure 15, two characteristics are common to most of the cases. Firstly, most of the 6000 uBqm" B e episodes exhibit an activity 3 7  concentration distribution starting at levels below 2000 u B q m before rising to values 3  exceeding 6000 uBqm" . 3  The Be activity concentration then remains at levels 7  exceeding 6000 uBqm" for a period of one to seven days before returning to sub 2000 3  uBqm" levels. 3  This indicates that stratospherically derived may have a substantial  influence over ground level ozone concentrations over a period of several days. Secondly, the C V of ozone is probably not a reliable measure of the spatial variability of ozone. In order to investigate the spatial and temporal variability of ozone, four urban stations and four rural stations were selected and their hourly ozone concentrations were plotted over the course of each episode. The stations used are listed in Table 5. Hourly ozone concentrations of four episodes corresponding to the episodes in Figure 15 are depicted in Figure 16. In Figure 16, each episode has two graphs. The first shows the hourly ground level ozone concentrations at the four urban stations and the second shows the hourly ground level ozone concentrations at the four rural stations.  38  a).  39  c).  26 January  28  26 January  1997  28  30  1997  80%ile 50%ile  \ /  y  26 January  28 1997  07 08 January  09 2001  30%ile  26  30  28  January  1997  07 08 January  09 2001  30  o 100 7 a  B  90  H  04  05  06  10  11  12  04  40  05  06  10  11  12  Figure 15 - Be activity concentration, average daily maximum temperature, average daily maximum ozone concentration and coefficient of variance of ozone for a) 25 October 1996 - 5 November 1996, b) 21 December 1996- 1 January 1997, c) 22 January 1997- 1 February 1997 and d) 4 January 2001 12 January 2001 from the Be only detection algorithm. A single vertical line represents an episode in which both Be and ozone were above the 80th percentile for one day only. Two vertical lines with an arrow between them indicate a multi-day episode. The graph on the upper left shows the daily Be activity concentration. The top horizontal line indicates the 80th percentile value. Values above this line constitute the top 20% of all daily Be activity concentrations in the data set. The lower line (marked GM) is the geometric mean daily Be activity concentration for the entire data set. The graph on the upper right shows the average daily maximum temperature, which is found by selecting and then averaging the daily maximum temperatures measured at both Abbotsford International Airport and Vancouver International Airport (Figure 11). For the graph on the lower left, the solid line is the average daily maximum ground level ozone concentration. The upper dotted line is the highest hourly ozone concentration recorded over the entire network on a given day. The lower dotted line is the lowest daily maximum value recorded over the entire network on a given day. The upper horizontal line indicates the 80th percentile value, the middle line represents the median value and the lower line represents the 30th percentile value. The graph on the lower right shows the coefficient of variation (CV) of ozone which is the standard deviation of the daily maximum ground level ozone concentrations recorded across the network divided by the average daily maximum ground level ozone concentration. 7  7  7  7  7  7  Urban  Station # T001 T002 TO 17 T021  Rural  Station Name  Station #  Downtown Vancouver Kitsilano Richmond South Port Coquitlam North  TO 12 TO 15 T027 T029  Station Name Chilliwack Airport Surrey East Langley Central Hope Airport  Table 5 - GVRD Air Quality Monitoring stations used in the hourly ozone concentration graphs in Figure 16. The locations of these stations can be found in Figure 11. Before analyzing the urban and rural graphs of episodes in which B e activity 7  concentrations exceeded 6000 uBqm" (Figure 16), it is instructive to compare them to 3  graphs of episodes in which the B e activity concentrations are low. Figure 17 shows 7  a high temperature low B e episode. 7  pronounced maximum  diurnal  The general pattern in such episodes is a  variability in the  concentration  usually  ground level  occurs  in the  ozone concentration. The  early  afternoon  hours  with  concentrations usually falling to below 10 ppb of ozone during the nocturnal hours, which is due to scavenging of ozone by N O (NRC 1992, Wayne, 1999). Figure 18 x  shows a low temperature episode, in which the B e activity concentrations are also 7  low. The general diurnal pattern is not as pronounced as the diurnal pattern in Figure  41  a). 90 85 80  Kitsilano Richmond South Port Coquitlam North Downtown Vancouver  Urban  75 70 65 60 ' 55 • 50 • 45 ' 40 • 35 • 30 ' 25 ' 20 ' 15 • 10 " 5 " 0 29  30 October 1996  90 85 80  Hope Airport Chilliwack Airport Surrey East Langley Central  Rural  75 70 65 60 55 50 '45 40 35 30 25 20 • 15 ' 10 • 5' 0 29  30 October 1996  November 1996  42  90  Kitsilano Richmond South Port Coquitlam North Downtown Vancouver  85 80 75 70 65 60 55 50 45 40 35 30 25 .I'':  20 15  h  M  10  a,  5 0  T 20  21  22  23  24  25  26  27  i 28  i  i"i 29  i 30  i  i' i 31  i 01  i 1 1 r ' i " i "i 1  02  03  December 1996  "T~l 22  I I I I 23  24  I  I  25  I  I  26  I I I I — I I I' I I  I  27  04  05  06  07  08  January 1997  28  29  December 1996  30  31  —I I V 01  02  I 03  I I" I ' i V "I"'I 1  04  January 1997  43  I T  1  05  06  07  08  90 85 80  Kitsilano Richmond Port Coquitlam North Downtown Vancouver  Urban  75 70 65 60 55 50 45 40 35 30 25 20 15 10 5 0 25 January 1997 90 85 80  Hope Airport Chilliwack Airport Surrey East Langley Central  75 70 65 60 55 50 45 40 35 30 25 20 15 10  4 «/  5 0 January 1997  44  60 H =5  1  50 H 45  1  40  "1  January 2001  45  Figure 16 (previous pages) - Hourly ozone concentrations measured at Kitsilano, Richmond South, Port Coquitlam North and Downtown Vancouver urban air quality monitoring stations (top graph) and Hope Airport, Chilliwack Airport, Surrey East and Langley Central rural air quality monitoring stations in the Greater Vancouver Regional District for a) 28 October 1996 - 1 November 1996, b) 20 December 1996 - 7 January 1997, c) 21 January 197 - 28 January 1997 and d) 6 January 2001 - 9 January 2001. Two vertical lines separated by an arrow indicate the duration during which the daily Be activity concentration remained above 6000 uBqm". 7  3  20  21  22  23  24  25 July  26  27  28  29  1996  Figure 17 - Hourly ozone concentrations measured Hope Airport, Chilliwack Airport, Surrey East and Langley Central rural air quality monitoring stations in the Greater Vancouver Regional District for 20 July 1996 - 29 July 1996.  17. This may be because very little ozone is produced from anthropogenic precursor pollutants as the ground level ozone concentration rarely rises above 20-25 ppb. A l l of the examples in Figure 16 have two distinct characteristics.  Firstly, all of  the urban ozone stations show very low ozone concentrations during the entire episode. This could be due to N O scavenging ozone in the urban areas. x  Secondly,  all of the rural areas exhibit peak ozone concentrations that do not rise above 40 ppb during any of the episodes, but, during most of the episodes, the concentration remains fairly constant and there seems to be very little nocturnal N O scavenging. x  46  90 -1 Hope Airport Chilliwack Airport Surrey East Langley Cenrat  65 80 75 " 70 65 " 60 "  S  55 -  a.  a ~  50 "  £  45 40 35 30 -  November  1996  Figure 18 - Hourly ozone concentrations measured Hope Airport, Chilliwack Airport, Surrey East and Langley Central rural air quality monitoring stations in the Greater Vancouver Regional District for 2 November 1998 - 11 November 1998. This characteristic is especially evident at Hope (T29 in Figure 11) and Chilliwack (T12) Airports and less so in Langley Central (T27) and Surrey East (T15).  3.3 Summary Figures 19-22 show the total number of days in each month during which both ozone and B e exceeded the 80 7  th  percentile (all data, exluding June - August and  excluding May - September) and during which the daily B e activity concentration 7  was greater than 6000 uBqm" respectively. 3  Upon inspection of the 80  th  percentile  graphs (Figures 19-21), April and May contain the greatest number of days during which both ozone and B e are in the 80 percentile. It may be tempting to perform 7  th  stratospheric ozone case studies during these particular months in Section 4. Unfortunately, these months do contain several days during which the average daily maximum temperature  exceeds 20°C.  Because of this, the influence of locally  47  Figure 19 - Total number of days in a month in which both the daily Be activity concentration and the average daily maximum ozone concentration exceeded the 80 percentile (all data) at the CTBTO IMS station CA002 during the period 9 April 1996- 12 March 2002. 7  th  Figure 20 - Total number of days in a month in which both the daily Be activity concentration and the average daily maximum ozone concentration exceeded the 80 percentile (excluding June - August) at the CTBTO IMS station CA002 during the period 9 April 1996 - 12 March 2002. 7  th  48  Figure 21 - Total number of days in a month in which both the daily Be activity concentration and the average daily maximum ozone concentration exceeded the 80 percentile (excluding May - September) at the CTBTO IMS station CA002 during the period 9 April 1996- 12 March 2002. 7  th  S  25  H  Figure 22 - Total number of days in a month in which the daily Be activity concentration exceeded 6000 uBqm' at the CTBTO IMS station CA002 during the period 9 April 1996- 12 March 2002. 7  3  49  produced anthropogenically derived ozone can not be ruled out.  Case studies in  Section 4 should therefore be carried out during months during which both the average daily maximum temperature is low (Figure 14) and the B e concentration is 7  high (Figure 22). Although the average daily maximum ozone concentration will be low during these periods, the low temperatures will ensure that a greater portion of the ground level ozone concentration in the L F V will have originated from natural, possibly stratospheric, sources.  4. Back Trajectories and Case Studies Meteorological trajectories are an important tool in stratosphere-troposphere exchange research (Stohl et al, 2000, 2002, Stohl and Trickl, 1999, Ancellet et al, 1991, Bonasoni et al, 2000, Danielsen, 1968). A trajectory is the time integration of the position of a parcel of air as it is transported by the wind (Stohl et al, 2002). The parcel's passive transport by the wind is computed from the average of the threedimensional velocity vectors at the particle's initial-position P(t) and its first-guess position P'(t+dt) ( R E A D Y website). The velocity vectors are interpolated in both space and time. The first guess position is:  P\t + dt) = P(t) + V(P,t)dt where V(P,t) is the velocity at the particle's initial position. The final position is:  \V(P,t) + V(P\t + dt)] P(t + dt) = P(t) + — ^ dt 1  K  J  Trajectories may be integrated both forward and backward in time ( R E A D Y website). Meteorological back trajectories are an important part of the current research because of their ability to determine the source altitude and location of near surface air in the L F V during episodes in which air of stratospheric origin is believed to be influencing ground level ozone concentrations.  50  Once back trajectories have been  calculated for a particular episode, meteorological analysis will be used to relate the trajectory paths to the overall synoptic situation. Early research utilizing meteorological trajectories required either an immense amount of computing power or the trajectories had to be computed by hand (Danielsen,  1968).  Trajectory calculations  were thus  considered  costly and  impractical for general use save for the largest research projects (Danielsen, 1968, Danielsen et al,  1970, 1977).  With the increase in the availability of fast desktop  PC's and workstations, trajectory analysis has become an important tool for many fields of atmospheric research such as stratosphere-troposphere exchange (Stohl et al, 2000, 2002, Stohl and Trickl, 1999, Ancellet et al,  1991, Bonasoni et al, 2000),  volcanic ash (Heffter and Stunder, 1993) and air pollution dispersion (Fast and Berkowitz, 1997, Gislason and Prahm, 1983). Several trajectory models are widely used in many fields of research. examples  are the F L E X T R A  Some  (FLEXible TRAjectory) model developed at the  Technical University of Munich in Germany (Stohl et al. 1995, Baumann and Stohl, 1997, Stohl and Seibert, 1998), the HYSPLIT (Hybrid Single Particle Lagrangian Integrated Trajectory) model developed by the N O A A Air Resources Laboratory in the U S A (Draxler and Hess, 1997) and the C A N E R M (CANadian Emergency Response Model) developed at the Canadian Meteorological Center in Dorval, Canada (Pudykiewicz, 1988, Pudykiewicz, 1989).  4.1 Previous Research The bulk of previous research involving trajectory analysis of S T E has been conducted in mountainous regions around the world, most importantly in the Swiss Alps because air of stratospheric origin has had little chance to dilute at these  51  altitudes. The following three papers, published over the last 12 years, outline some of the previous work. The study conducted by Ancellet et al. (1991) used a ground-based lidar at the Observatoire de Haute Provence (OHP) in France to determine a coarse value of the total outflow of ozone from the stratosphere to the troposphere due to frontal activity. Trajectory analyses were used for several cases to determine whether the elevated ozone concentrations measured at the observatory were the result of S T E . For each model run, trajectories were calculated at end points both directly overhead and 1° north, south, east and west of the observatory. European  Center for  Medium  Range  The model input consisted of  Weather Forecasting ( E C M W F )  dimensional mean wind fields at a resolution of 2° x 2°.  three  Using these trajectory  calculations for an elevated ozone case that occurred on 28-29 March 1988, it was found that surface air at OHP originated from a tropopause fold northeast of England two days earlier. Using the ozone lidar at OHP, Ancellet et al. (1991) estimated that a total of 2 x l 0  33  ozone molecules were transferred via the tropopause fold.  This  number was multiplied by the average annual number of tropopause folds to give an approximate value for the total ozone outflow from the stratosphere in the Northern Hemisphere (1.5xl0  36  molecules yr" ).  from previous studies (3.5-6.1xl0  1  36  This number is lower than values obtained  molecule yr" ) because this study did not include 1  the ozone outflow generated by cutoff lows (Danielsen and Mohnen, 1977). Bonasoni et al. (2000) used a variety of meteorological data including trajectory analysis, water vapor satellite imagery, ozone soundings, ground based air quality monitoring and synoptic charts to describe a stratospheric intrusion event at Mt. Cimone (2165 m a.s.l.) in the Northern Apennines in Northern Italy. The case study took place on 11-12 March 1996 during the Vertical Ozone Transport in the Alps  52  ( V O T A L P ) field study.  Around midday on 12 March an ozone concentration of  approximately 1 lOppb was measured at Mt. Cimone. A n ensemble of trajectory end points situated around the Mt. Cimone area was set up using the F L E X T R A trajectory model to determine the source region of this ozone rich air. The bulk of the resulting trajectory pathways showed that the air around the Mt. Cimone area originated near Greenland seven days earlier. This air then became incorporated into a cut off low over Central Europe prior to its arrival over Mt. Cimone. Thirty six hours earlier the trajectory paths originated at low levels over Northwestern Italy, suggesting an anthropogenic source. During this case study, trajectory analysis was essential as it was able to rule out ozone of anthropogenic origin during the elevated ozone event. Another case study that occurred during the V O T A L P project was investigated by Stohl and Trickl (1999). March  1997  ozone  Between midday on 28 March and early evening on 29  concentrations  between 60-80ppb  were  observed  at the  Jungfraujoch observatory in the Swiss Alps. During this time, the relative humidity (RH) fell to nearly 10% at Jungfraujoch starting at midday on 28 March and persisted until midday on 29 March. These low R H values suggested that air of stratospheric origin was heavily influencing the ground level ozone concentrations. After midday on 29 March, the R H rose to about 90%, yet the ozone concentrations remained elevated thus suggesting a possible anthropogenic influence on ground level ozone concentrations. In order to investigate the source regions of these two air masses, the F L E X T R A trajectory model was utilized, using E C M W F 1° x 1° mean wind fields. After analyzing the trajectory results, Stohl and Trickl (1999) concluded that ozone concentrations measured at Jungfraujoch during the period of low relative humidity originated from a S T E event over Northern Europe. Ozone concentrations measured  53  during the high relative humidity period were the result of long range transport of air that originated in the boundary layer over Eastern North America.  4.2 Trajectory Models and Procedures Used in the Current Research 4.2.1 HYSPLIT The N O A A A i r Resource Laboratory's (ARL) Hybrid Single Particle Lagrangian Integrated Trajectory (HYSPLIT) model is designed to support a wide range of simulations including long range transport (trajectories), particle dispersion and deposition.  The model input into HYSPLIT is based on the N O A A Global Data  Assimilation System (GDAS) three dimensional gridded wind fields.  The G D A S  model is run every six hours with start times at 00, 06, 12 and 18 U T C at one degree latitude-longitude resolution.  A R L then interpolates the G D A S wind fields to a  conformal map projection which reduces the resolution to 100km.  The resulting  modified G D A S model output is called ARL-packed F N L output. F N L output files are available for both the Northern and Southern Hemispheres. contains two weeks worth of compacted G D A S model output.  Each F N L file  The resulting F N L  files are used as the model input for HYSPLIT.  4.2.2 CANERM The capabilities of the Canadian Meteorological Center's ( C M C ) CANadian Emergency Response ( C A N E R M ) model are similar to H Y S P L I T .  The model runs  performed on C A N E R M used three dimensional mean wind fields from the N C E P (National Center for Environmental Prediction) Reanalysis model output with a latitude-longitude grid resolution of 1° degree.  The N C E P reanalysis is available  every six hours at 00, 06, 12 and 18 U T C . The N C E P reanalysis output is based on a project to recover archived data such as land surface, ship, rawinsonde, pibal, aircraft, satellite and other data. These data are then quality controlled and assimilated with a  54  data assimilation system, which is kept unchanged over the reanalysis period. This eliminates perceived climate jumps associated with changes in the data assimilation system. Further information on the N C E P reanalysis project can be found in Kalnay  etal. (1996).  4.2.3 Trajectory Model Procedure Twenty seven trajectories were calculated around the L F V area for each model run during an episode. O f these 27 trajectories, nine were calculated starting at 100m a.g.l. (above ground level), nine were calculated starting at 500m a.g.l. and nine were calculated starting at 1000m a.g.l. The nine end point receptors at each of the three levels are identical. Eight of these endpoints are evenly distributed around the edge of a l ° x l ° square centered around 4 9 . 2 5 ° N latitude 123.25°W longitude.  The ninth  endpoint receptor is located at the center of the l ° x l ° square. The total duration of each back trajectory is 144 hours and the position of each trajectory is recalculated every hour.  4.3 Case Studies In order to better understand the meteorological phenomena that occur both prior to, and during an elevated ground level ozone concentration episode in the L F V that is believed to have been the result of STE, a set of four episodes were selected for more detailed study. available),  During each case study, synoptic analysis, satellite imagery (where  temperature  and  precipitation,  back  trajectory  analysis,  ozone  concentration and B e activity concentration data were utilized. These case studies 7  were not selected randomly from the identified episodes. Instead, they were selected based on a detailed set of criteria listed below. 1.  7  B e exceeds a threshold activity concentration during the episode.  The  selection of the threshold activity concentration was based on previous studies.  55  Dutkiewicz and Husain (1979) suggested that a 24 hour ground level B e 7  activity of 6000 uBqm" , roughly one third of upper tropospheric values, 3  indicates a recent upper atmospheric contribution of B e . 7  2.  Ozone concentrations remain elevated during most of the episode, especially during the night time hours.  3.  The average daily maximum temperature must be below 20°C (see section 3.1.1).  4.  A majority of the back trajectories from any one model run show a total descent of at least 4000 m - 5000 m over a six day period immediately prior to the trajectory's arrival in the L F V (Stohl et al, 2000). This last criterion is implemented to ensure that lower tropospheric sources of ozone are kept to a minimum.  It is entirely possible that ozone originating from 5000 m a.s.l.  (above sea level) is of anthropogenic origin (i.e. long range transport). If the air mass at that altitude exhibits a high B e activity concentration however, the 7  possibility that the air mass is of stratospheric origin increases. 7  Even if the  B e activity concentration at 5000 m a.s.l. is high, there is still a small chance  that the air mass could still have an upper tropospheric origin because this region constitutes 25% of the B e production. During the case study selection 7  process, preference was given to episodes in which the bulk of the trajectory descent took place three to four days immediately prior to the trajectory's arrival in the L F V thus minimizing the influence of anthropogenically derived ozone along the trajectory path.  56  4.3.1 Case Study #1 - 28 October 1996-31 October 1996 4.3.1.1 Synoptic Overview On 23 October 1996 at 0000 U T C , a large scale trough was present over most of the Northern Pacific Ocean. The trough axis was oriented northwest - southeast with its most intense portion (i.e. area of strongest gradient of geopotential height contours) located between 3 0 ° N - 4 0 ° N latitude and 170°E longitude. In addition, a cut off low was located in the Arctic Ocean north of Alaska, U S A with its center located at approximately 7 5 ° N latitude and 140°W longitude. Both features were present at least up to the 250 mb isobaric surface. Over the next 36 hours, the wavelength of the Pacific trough began to decrease. At the same time, the amplitude of the trough began to increase and its axis changed to a more north - south orientation. At the same time, the cutoff low in the Arctic Ocean began to slowly track southwestward so that its center of rotation was located just to the north of Alaska. On 25 October at 1200 U T C , the flow around the Pacific trough began to separate from the main flow allowing a cutoff low to develop at 500 mb. The cutoff low was not apparent at 250 mb. Immediately east of the cutoff low, a steep ridge began to develop over the northeastern Pacific Ocean and a trough over western North America began to dig southwards. The cutoff low centered over Northern Alaska did not move relative to its position 18 hours previously. These synoptic patterns on the 250 mb and 500 mb isobaric surfaces are shown in Figure 23 and Figure 24 respectively. Over the next 72 hours, the Pacific cutoff low rejoined the main flow and the associated trough began to weaken. At the same time, the Northern Alaska cutoff low began to track eastwards so that its center was located just north of the Northwest Territories in Canada.  57  At the surface, several main synoptic processes took place.  On 23 October at  0000 U T C a low was present over the central Alutian Islands located west of Alaska in the U S A with a central pressure of about 984 mb. Over the next 72 hours, the low began to fill and track eastwards towards the Alaskan Panhandle. Associated with the Pacific trough present at 0000 U T C on 23 October, a new low began to develop on the east side of the trough with a central pressure of 1011 mb. Over the next 48 hours, the low moved towards the east and deepened to 969 mb so that its center was now situated just south of the central Aleutian Islands. At the same time, the previous low that began to dissipate over the Alaskan Panhandle was incorporated into this low. After 25 October at 1800 U T C , this low began to track further east and fill as it came onshore over northern British Columbia. During the lifetime of this low, a large area of high pressure extending from the southeastern Pacific to Western North America began to develop. Besides a short lived low that developed over Southwestern British Columbia on 28 October at 1800 U T C , this high pressure pattern persisted for the duration of the elevated ozone episode in the L F V . The final low pressure system to track over the Northern Pacific was the most intense with central pressures reaching 956 mb.  The low developed over eastern  Siberia on 24 October at 1800 U T C and over the next 72 hours it moved northeastward to northeast Russia where it reached its lowest central pressure (Figure 25). On 28 March at 0000 U T C , the low began to track further to the northeast where it began to steadily dissipate over the northern entrance to the Bering Straight.  4.3.1.2 Surface O3 and Be Concentrations and Temperature in the LFV 7  On 28 October at 0600 U T C (27 October at 2200 Pacific Standard Time), elevated ozone concentrations began to be measured in the L F V starting at Hope Airport (Figure 16a). Over  the  next  12 hours elevated  58  ozone  concentrations  were  59  60  61  Figure 25 (previous page) - Surface isobaric chart for 28 October 1996 at 0000 UTC. Image courtesy of Environment Canada's Canadian Meteorological Center.  measured at Chilliwack Airport, Surrey East and Langley Central air quality stations. With the exception of Surrey East, these ozone concentrations remained at a roughly constant level during the entire five day episode with very little N O  scavenging  x  during the night time hours. The maximum 1 hour ground level ozone concentration measured at each rural station in Figure 16a during the episode ranged between 25-40 ppb. At the urban stations, very little evidence of elevated ozone concentrations could be found (Figure 16a). The only exception occurred on 29 October at 0600 U T C (28 October at 2200 PST) when a small spike between 10-20 ppb was measured at the Kitsilano and Richmond South air quality stations. During the episode, the average daily maximum (ADM) ozone concentration rose by 10 ppb from a pre-episode concentration of 14 ppb on 27 October to 24 ppb on 2930 October (Figure 15a). The A D M ozone concentration deceased to 17 ppb by the end of the episode on 31 October.  The coefficient of variance (CV) of ozone  experienced a relative minimum during the entire episode with values ranging from 0.4 ppb between 28-30 October to 0.5 ppb on 31 October (Figure 15a). This relative minimum in C V of ozone signifies that the maximum daily ozone concentrations across the network were approaching a more common value compared to the maximum daily ozone concentrations immediately prior to and following the episode. Although ozone concentrations remained elevated for four days, the B e activity 7  concentrations remained above 6000 uBqm" for two days only (30-31 October) 3  (Figure 15a).  A possible explanation for this would be that a heavy precipitation  event occurred on 28 October. Vancouver International Airport (Figure 11) measured  62  50 mm of total precipitation on this day.  Since B e can be scavenged out of the 7  atmosphere by precipitation (Feely et al, 1989), it is likely that a large amount of B e 7  was scavenged out of the atmosphere above the L F V . Over the course of the entire episode, the average daily maximum temperature remained fairly constant, ranging from 10°C-11°C.  4.3.1.3 Back Trajectory Analysis Back trajectory analysis, performed between 30-31 October, when the daily B e 7  activity concentration measured at U B C exceeded 6000 uBqm" , revealed two major 3  trajectory paths that air particles followed prior to their arrival over the L F V . Figure 26 shows six day back trajectories taken by particles of air that arrived over the L F V on 31 October at 0600 U T C .  The first trajectory path followed the flow around the  Pacific trough and originated as far away as North Central Siberia six days prior (Figure 26). A l l of the trajectories that arrived over the L F V at an altitude of 1000 m a.g.l. that traveled in the flow around this trough originated at a height of at least 5000 m a.g.l. Prior to their arrival in the L F V , these trajectory paths descended at a rate of 1.5-2.0 cms" between 27 October at 0600 U T C and 30 October at 0600 U T C . This 1  rapid descent was probably due to the large surface high pressure region present over Alaska and the northern two thirds of British Columbia. With the exception of four trajectory paths which originated at low levels over the northeastern Pacific, the remaining trajectories originated around the region of the North Alaskan cutoff low six days earlier.  After exiting the cutoff low, these  trajectories descended at a rate of approximately 2.0 cms" over three days before 1  arriving over the L F V (Figure 26). Again, this rapid descent was probably caused by the two surface high pressure systems present over Alaska and British Columbia. These two major trajectory paths persisted for the entire period during which ozone  63  160°E  -144  -132  -120  -106  -96  -84  -72  -80  -48  -36  -24  -12  0  Tirre (hours) Figure 26 - Three dimensional back trajectories arriving over the L F V area on 31 October 1996 at 0600 U T C using the C A N E R M model. The three day back trajectory ensemble consists of 27 trajectories starting at 100m a.s.l., 500m a.s.l. and 1000m a.s.l. over the L F V . The lower plot shows the time height profile o f the trajectories. Back trajectory data courtesy of Environment Canada's Canadian Meteorological Center.  concentrations remained elevated and daily Be activity concentrations exceeded 6000 7  uBqm" . 3  Even though neither of the two  major trajectory paths originated in the  stratosphere, it is still probable that these trajectories passed through high ozone concentration regions prior to their arrival in the L F V .  It has long been known that  cutoff low pressure systems and troughs are responsible for the descent of ozone rich stratospheric air into the mid to upper troposphere via tropopause folding (Ancellet et al.,  1993, Johnson and Viezee, 1981, Haagenson, 1981, Danielson et al., 1970,  Schuepbach et al., 1998, Davies and Schuepbach, 1993).  Therefore, air need only  pass through these regions at mid to upper tropospheric levels to intercept ozone rich,  64  stratospheric air. This possible scenario is supported by the elevated ground level ozone and B e concentrations at the surface in the L F V . 7  4.3.2 Case Study #2 - 20 December 1996- 01 January 1997 4.3.2.1 Synoptic Overview On 21 December 1996 at 0000 U T C , three main upper level features were present over the northeastern Pacific and North America. high amplitude trough was present.  Over the north central Pacific, a  The trough axis was oriented in a northwest -  southeast direction. Immediately to the east of the trough, a narrow, high amplitude ridge extended as far north as the Arctic Ocean with its axis tilted in a north - south direction.  The final feature was a broad trough encompassing the entire North  American continent.  These features were present at least up to the 250 mb isobaric  surface. Over the next 24 hours, the Pacific ridge and trough detached from the main flow creating a cutoff high and cutoff low respectively.  Between 23 December at 0000  U T C and 24 December at 0000 U T C , the cutoff low moved eastward while the cutoff high, centered over northwestern Alaska remained stationary. This caused the cutoff high to become completely separated from the main flow (Figures 27 and 28). During this 24 hour period, the geopotential gradient of the large scale trough over North America increased and its axis shifted from the west coast of North America to the central United States. Over the next 5 days, both the Pacific cutoff low and cutoff high joined and then re-detached from the main flow.  The major consequence of this ridge/cutoff high  system was that it acted as a barrier to the further eastward propagation of many surface low pressure systems that formed in the western and central Pacific Ocean. During this same time span, the northern portion of the North American trough  65  99  L9  separated from the main flow to create a cutoff low that covered most of Central Canada and the Northwest Territories. At the surface, four major synoptic patterns were evident. These synoptic systems corresponded well with the upper level flow described above.  Associated with the  Pacific upper level trough on 21 December at 0000 U T C were three surface low pressure systems with the northern most system being the most intense with a central pressure 972 mb. Separating these low pressure systems from the west coast of North America was a strong high pressure system with a central pressure of 1048 mb over the northern Yukon Territory, Canada.  Over North America, this high pressure  system covered most of central Canada and separated two weak surface lows over the northern United States from an intense low over Labrador, Canada with a central pressure of 954 mb. As the upper level ridge and trough began to separate from the main flow on 22 December at 0000 U T C , the North American high pressure system began to intensify and extend further south. Because of the strong pressure gradient on the western side of this high, the eastward progression of several surface low pressure systems over the northern Pacific was hindered. Over the next six days, these blocked low pressure systems combined to form one large surface low with several frontal systems (Figure 29). Associated with the large upper level cutoff low that began to form over northern Canada on 24 December at 0000 U T C were several surface low pressure systems centered over an area just to the west of Baffin Island in the Northwest Territories. This area of low pressure was unable to progress further southwards due to the large area of high pressure over central Canada (Figure 29).  68  69  Figure 29 - Surface isobaric chart for 28 December 1996 at 1800 UTC. Environment Canada's Canadian Meteorological Center.  Image courtesy of  4.3.2.2 Surface O3 and Be Concentrations and Temperature in the LFV 7  On 21 December at 0600 U T C (20 December at 2200 PST), elevated ground level ozone concentrations began to be measured in the L F V .  Unlike Case #1 described  above (Section 4.3.1.2), these elevated concentrations occurred simultaneously at all four rural air quality monitoring stations (Figure 16b). By 21 December at 1200 U T C (0400 PST), the ozone concentrations at the rural stations rose to 25-30 ppb. These ozone concentrations rose to 30-35 ppb over the next seven days with very little scavenging by N O during the night time hours. After 29 October at 2200 U T C (1600 x  PST), the ground level ozone concentrations began to decrease until 31 December at 0600 U T C (30 December at 2200 PST) when concentrations began to sharply increase again with concentrations rising to 40-45 ppb. These concentrations did not remain as constant as the concentrations measured over the prior 10 days. The urban air quality monitoring stations experienced ground level  ozone  concentrations with more variability at a diurnal scale especially at Kitsilano and Downtown Vancouver (Figure 16b). During the night time hours however, elevated ozone concentrations were measured at both the Richmond South and Port Coquitlam North air quality stations with concentrations reaching 30 ppb. During the episode, the A D M ozone concentration remained fairly constant at about 20 ppb (Figure 15b).  The only exceptions occurred when the A D M ozone  concentration dipped to 16 ppb on 24 December and then rose to 27 ppb on 25 December. The C V of ozone also remained fairly constant during the episode (0.4 ppb) except between 24 December - 26 December (Figure 15b). During this time, the  70  C V rose to 0.6 ppb on 24 December and then fell to 0.2 ppb on the 25 December which corresponded with the relative maximum in A D M ozone. As in Case #1, the daily B e activity concentrations 7  uBqm"  3  remained  above  6000  for two days only even though the ground level ozone concentrations  remained elevated for several days (Figure 15b).  Precipitation may have been  responsible for the low activity concentrations measured at the beginning of the episode. On 20-21 December, the daily total precipitation at Vancouver International Airport was 7.9 mm and 8.5 mm respectively.  Ishikawa et al. (1995) showed that  even light precipitation events could washout a large amount of Be. 7  The A D M temperature fluctuated between a low of -5°C on 23 December to a high of 6°C on 29 December. occurred between 21  The relative minimum in A D M temperature that  December -  26 December could have been due to a  cold/stationary front present over the area at this time.  4.3.2.3 Back Trajectory Analysis When the daily B e activity concentration exceeded 6000 uBqm" on 26-27 7  3  December, back trajectory analysis revealed three major trajectory paths that air particles followed prior to their arrival over the Lower Fraser Valley (Figure 30). The first trajectory path originated in the Pacific trough at levels just below the tropopause. Insufficient data was available to conclude whether this air originated in a tropopause fold associated with the upper level trough or if the air had just passed through an upper tropospheric region in which stratospherically intruded air was present.  These trajectories originated from a maximum altitude of 7200 m before  descending into the mid troposphere as the air traveled around the ridge and then to the lower troposphere as the air traveled through the broad region of surface high pressure that was located over British Columbia (Figure 29).  71  -144  -132  -120  -106  -96  -84  -72  -60  -43  -36  -24  -12  0  Time (reus)  Figure 30 - Three dimensional back trajectories arriving over the LFV area on 26 December 1996 at 1200 UTC using the CANERM model. The three day back trajectory ensemble consists of 27 trajectories starting at 100m a.s.l., 500m a.s.l. and 1000m a.s.l. over the LFV. The lower plot shows the time height profile of the trajectories. Back trajectory data courtesy of Environment Canada's Canadian Meteorological Center.  The next trajectory path originated in the lower troposphere over the eastern United States.  This air then traveled up the eastern seaboard of Canada where it  ascended into the cutoff low situated over northern Canada on 21 October at 1800 U T C . The air then rapidly descended into the lower troposphere at an average rate of 2.3 cms" as it exited the cutoff low on 24 December at 0000 U T C and encountered 1  the area of strong surface high pressure over northern Canada. The final trajectory path originated in the lower troposphere and circulated around a surface low pressure system off the coast of southern British Columbia. The bulk of the trajectories, whose end points originated between 100 m a.s.l. and 500 m a.s.l. over the L F V region, followed this trajectory (Figure 30).  72  It is entirely possible that the trajectories that originated over eastern North America could have contained ozone from two completely different sources: the stratosphere and the boundary layer. The end of the back trajectory paths originated in the free troposphere at heights between 1500-2500 m.  This air could have  contained photochemically derived ozone produced over the highly urbanized American east coast. As this air ascended into the northern Canadian cutoff low, it may have encountered ozone rich air that descended from the stratosphere. Stohl and Trickl (1999), using detailed back trajectory analysis, studied a similar situation that occurred in the Swiss Alps when air arriving at the Jungfraujoch observatory contained ozone that originated within the boundary layer over the eastern United States and a tropopause fold that occurred over Greenland.  4.3.3 Case Study #3 - 22 January 1997 - 29 January 1997 4.3.3.1 Synoptic Overview On 18 January at 0000 U T C , three large troughs were present over the northern Pacific Ocean and North America.  The first trough, centered over the Russian  Kamchatka Peninsula, covered the entire western half of the Pacific. Its trough axis was orientated in an approximately north - south direction. Over the next 24 hours, the northern section of the trough axis shifted to a northwest - southeast orientation. At the same time, this section of the trough also began to separate from the main flow forming a cutoff low (Figure 31). Between 19 January at 0000 U T C and 22 January at 1200 U T C , the southern portion of this trough began to dig southeastwards and formed another cutoff low centered about 3000 km north of Hawaii. These two cutoff lows intensified before rejoining the main flow 72 hours later.  Separating this  western Pacific trough/cutoff low system was a ridge that began to extend northwards  73  over the Bering Straight on 18 January at 0000 U T C before forming a cutoff high on 25 January at 0600 U T C . The second trough present over the eastern portion of the north Pacific, on 18 January at 0000 U T C with an approximately north - south oriented trough axis, was short lived (Figure 31). Over the next 48 hours, the wavelength of this trough began to decrease as it propagated eastward so that its axis was now situated near the west coast of North America. Between 20 January at 0000 U T C and 23 January at 0000 U T C , the trough began to weaken and finally dissipate as it crossed the western North American coast. The third very broad trough existed over the eastern half of North American continent for the duration of this episode (Figure 31).  In the northern part of the  trough, a cutoff low, with a large geopotential gradient, was present over central Quebec, Canada.  Over the next five days, this cutoff low began to propagate  northwestwards and weaken with its new position centered over Baffin Island in northern Canada. Between 21 January at 0600 U T C and 24 January at 1800 U T C , the southern part of the trough began to move westwards with its axis now centered over central North America. At the same time, the trough began to dig so that it extended as far south as the central portion of the Gulf of Mexico. A l l three of these troughs were present at least up to the 250 mb isobaric surface. At  the  surface,  there  were  up to  five  low  pressure  systems occurring  simultaneously associated with the upper level trough present over the Western Pacific (Figure 32). Because of the surface high pressure system, associated with the central Pacific upper level ridge (Figure 31), centered over the northeastern Pacific Ocean, these lows were not able to propagate further east. As this high pinched off from a much larger region of high pressure over the tropics on 21 January at 0000  74  U T C , the west Pacific low pressure systems were able to move eastwards around the southern portion of this "pinched o f f high. These lows then encountered another region of high pressure centered over central Canada, blocking their further eastward progression. This was due to the strong pressure gradient associated with the outer region of this high (Figure 33). These lows eventually dissipated on 24 January. As these lows tracked around the southern end of the "pinched o f f high pressure system and began to propagate and dissipate over the southeast portion of the northern Pacific and the southwestern United States, a new, more intense set of surface lows began to form in the north central Pacific. The movement of these lows was also restricted due to the strong surface high pressure system present over northwestern Canada. The low pressure systems associated with the short lived trough present over the northeastern Pacific began to weaken between 18 January at 0000 U T C and 19 January at 0000 U T C . These lows eventually dissipated completely on 25 January at 0000 U T C . The final set of low pressure systems were associated with the cutoff low present over northeastern Canada.  Starting on 21 January at 0000 U T C , two new low  pressure systems began to form over south central Canada. Over the next three days, these lows tracked eastwards over the Great Lakes, before turning north and joining the set of low pressure systems associated with the cutoff low over northeastern Canada.  76  77  78  Figure 32 (page 77) - Surface isobaric chart for 18 January 1997 at 1800 UTC. Image courtesy of Environment Canada's Canadian Meteorological Center. Figure 33 (previous page) - Surface isobaric chart for 25 January 1997 at 1800 UTC. Image courtesy of Environment Canada's Canadian Meteorological Center.  4.3.3.2 Surface O3 and Be Concentrations and Temperature in the LFV 7  Between 22 January at 1200 U T C (0400 PST) and 23 January at 0000 U T C (22 January at 1600 PST), elevated ozone concentrations began to be measured at the rural air quality monitoring stations in the L F V (Figure 16c). With the exception of the Hope Airport air quality station, the ozone concentrations exhibited a diurnal pattern between 23 January at 0000 U T C (22 January at 1600 PST) and 25 January at 0800 U T C (0000 PST). Briefly, a normal diurnal pattern of ozone exhibits a morning low in concentration rising to maximum levels in the early afternoon and then decreasing to low nocturnal concentrations around mid-late evening (Figure 12). For the next 72 hours, the elevated ozone concentrations remained constant between 2030 ppb with no N O scavenging during the night time hours (with the exception of the x  Surrey East station).  After 28 January at 0800 U T C (0000 PST), the ozone  concentrations began to decrease and exhibit a more diurnal pattern. At the beginning of the episode, urban ozone concentrations rose sharply in the Kitsilano and Richmond air quality stations from near zero levels to between 15-30 ppb. The Port Coquitlam North and Downtown Vancouver stations did not exhibit a similar increase until two days later (Figure 16c).  At about the same time in which  the rural ozone concentrations switched to a constant ozone concentration pattern from a diurnal pattern, the urban ozone concentrations sharply fell to near zero values on 25 January at 0000 U T C (24 January at 1600 PST). With the exception of the night time hours on 26 January and 27 January, the ozone concentrations exhibited a  79  roughly diurnal pattern between 25 January at 0000 U T C (24 January at 1600 PST) and the end of the episode on 28 January. During the episode, the A D M ozone concentration fluctuated between a maximum of 30 ppb on 23 January and a minimum of 17 ppb on 27 January (Figure 15c). The maximum A D M ozone concentration corresponded to the time at which the hourly ozone concentrations began to sharply increase in the urban and rural stations (Figure 16c). Strangely enough, the minimum A D M ozone concentration corresponded with the maximum daily B e measurement taken at U B C . 7  The C V of ozone reached a  minimum of 0.15 ppb at the same time that the A D M ozone concentration reached its maximum and vice versa for the minimum C V of ozone (Figure 15c). Daily B e activity concentrations measured at the beginning of the episode (Figure 7  15c) were low even though ground level ozone concentrations were elevated at this time. Precipitation may have been responsible for these low activity concentrations as total daily accumulations of between  1.5-30 mm were measured at Vancouver  International Airport between 16-22 January. The A D M temperature decreased from 7°C on 22 January to a minimum of -2.9°C on 25 January. Cold easterly outflow winds from the continent behind a stationary front located over the Lower Fraser Valley area may have been responsible for the cold temperatures that occurred on 25 January.  A D M temperatures increased to  approximately 4°C by the end of the episode on 28 January.  4.3.3.3 Back Trajectory Analysis Back trajectory analysis revealed that the initial elevated ozone concentrations measured in the L F V may have been the result of two sources.  The first source  originated either within or just above the boundary layer over the west coast of the United States (Figure 34). As the air associated with this trajectory path traveled up  80  -144  -132  -120  -106  -83  -84  -72  -60  48  -36  -24  -12  0  Time (hours)  Figure 34 - Three dimensional back trajectories arriving over the L F V area on 23 January 1997 at 0000 UTC using the CANERM model. The three day back trajectory ensemble consists of 27 trajectories starting at 100m a.s.l., 500m a.s.l. and 1000m a.s.l. over the LFV. The lower plot shows the time height profile of the trajectories. Back trajectory data courtesy of Environment Canada's Canadian Meteorological Center.  the coast, it may have encountered polluted boundary layer air from many large urban costal centers including San Diego, Los Angeles, San Francisco, Portland and Seattle. The second trajectory path originated at mid tropospheric levels within a cutoff low centered just to the south of Alaska (Figure 35). Starting on 24 January at 1800 U T C , the bulk of the trajectories originated at mid tropospheric to lower stratospheric levels in the Arctic region between northern Alaska and north central Siberia (Figure 36). The trajectories that descended from the highest altitudes originated on the eastern side of a trough located over central Siberia six days earlier. The remaining high altitude trajectories originated from a shortwave trough that formed over northeastern Russia on 20 January at 0600 U T C . On 23  81  Time (hours)  Figure 35 - Three dimensional back trajectories arriving over the LFV area on 23 January 1997 at 1800 UTC using the CANERM model. The three day back trajectory ensemble consists of 27 trajectories starting at 100m a.s.l., 500m a.s.l. and 1000m a.s.l. over the LFV. The lower plot shows the time height profile of the trajectories. Back trajectory data courtesy of Environment Canada's Canadian Meteorological Center.  January at 0000 U T C , the bulk of the high altitude trajectories began to descend into the mid to lower troposphere at an average rate of 1.2-2.2 cms" . This descent was 1  probably caused by the large region of high pressure that was present over western Canada and Alaska. Since the tropopause can exist at heights as low as 8000 m above sea level in the polar regions (Stull, 2000), it is entirely possible that the trajectories that originated at around 8000 m in the Siberian trough could have originated in the stratosphere. Similar to Case #1, it is possible that two sources of stratospheric air, both from the Siberian trough and the shortwave located over northeastern Russia, contributed to the  82  Figure 36 - Three dimensional back trajectories arriving over the LFV area on 26 January 1997 at 1200 UTC using the CANERM model. The three day back trajectory ensemble consists of 27 trajectories starting at 100m a.s.l., 500m a.s.l. and 1000m a.s.l. over the LFV. The lower plot shows the time height profile of the trajectories. Back trajectory data courtesy of Environment Canada's Canadian Meteorological Center.  elevated ground level ozone concentrations in the L F V between 25 January at 0800 U T C and 28 January at 0800 U T C . As in Case #2, the elevated ozone concentrations in the L F V at the beginning of the episode may have been both stratospheric (south Alaskan cutoff low) and lower tropospheric (boundary layer over the west coast of North America.) in origin (Figures 34 and 35).  4.3.4 Case Study #4 - 05 January 2001 -11 January 2001 4.3.4.1 Synoptic Overview On 1 January at 0000 U T C , two troughs were present over the northern Pacific Ocean.  The westernmost trough was centered over Japan with an approximately  83  north - south oriented axis. The second trough covered the entire northeast Pacific, with its  axis centered at about 155°W longitude with a north - south orientation.  Separating these two troughs was a ridge that extended as far north as northeastern Russia (Figure 37). The upper air pattern over North America consisted of a ridge extending as far east as Western Ontario, Canada with its axis oriented in a northeast - southwest direction. To the east of this ridge was a trough that extended as far east as Saint John's, Canada with is axis centered over Texas, U S A to the south and the Great Lakes to the north (Figure 37). Over the next 48 hours, the ridge situated over the central Pacific began to separate from the main flow allowing a cutoff high to form on 3 January at 0000 UTC.  This cutoff high began to propagate in a northwesterly direction until it  rejoined the flow over north central Siberia on 6 January at 0000 U T C . As the central Pacific ridge began to separate from the main flow, the two Pacific troughs combined to form one large trough that covered the entire north Pacific on 2 January at 1200 U T C .  Extending over the entire trough was a jet streak situated at  250 mb in which wind speeds exceeded 150 knots (Figure 38).  Over the next six  days, a number of shortwave troughs propagated around this Pacific trough, the largest of which occurred on 7 January at 1200 U T C (Figure 39).  Between 31  December and 8 January, the North American ridge to the west and trough to the east persisted at various degrees of intensity. Over the course of the entire episode, the synoptic features corresponded well with the upper air flow. surface low  pressure  Associated with the western Pacific trough was an intense  system centered  approximately  Japan with a  84  800 km  east of central  85  Figure 37 (previous page) - 500 mb geopotential chart for 1 January 2001 at 0000 UTC. Image courtesy of Environment Canada's Canadian Meteorological Center.  central pressure of 965 mb on 1 January at 0000 U T C (Figure 40). Over the next 24 hours, this surface low tracked northward towards the Kamchatka Peninsula and deepened to 950 mb. To the east of this low was a large region of high pressure located over the central Pacific (Figure 40). As the corresponding ridge began to separate from the main flow on 1 January at 1200 U T C , this surface high pressure system retreated to the southeast allowing the low pressure systems in the east and west Pacific to join and form a region of surface low pressure covering the entire north Pacific. The northeastern Pacific was dominated by five surface low pressure systems with central pressures ranging form 979 mb-1000 mb. Most of these low pressure systems were unable to make landfall over the west coast due a large region of high pressure covering the western half of North America. These low pressure centers joined with the west Pacific surface lows as the central Pacific surface high retreated to the southeast (Figure 41). At any one time, the number of lows associated with this large Pacific surface low pressure region ranged from four to eight. As the region of high pressure over western North America began to weaken on 8 January at 1200 U T C , the easternmost low pressure centers made landfall over northwestern Canada where they rapidly dissipated.  86  87  88  89  90  Figure 38 (page 87) - 250 mb geopotential chart for 2 January 2001 at 1200 UTC. Image courtesy of Environment Canada's Canadian Meteorological Center. Figure 39 (page 88) - 500 mb geopotential chart for 7 January 2001 at 1200 UTC. Image courtesy of Environment Canada's Canadian Meteorological Center. Figure 40 (page 89) - Surface isobaric chart for 1 January 2001 at 0000 UTC. Image courtesy of Environment Canada's Canadian Meteorological Center. Figure 41 (previous page) - Surface isobaric chart for 6 January 2001 at 1200 UTC. Image courtesy of Environment Canada's Canadian Meteorological Center.  4.3.4.2 Surface O3 and Be Concentrations and Temperature in the LFV 7  On 6 January between 0000-1000 U T C (5 January at 1600 PST to 6 January at 0200 PST), the rural ground level ozone concentrations exhibited a diurnal pattern at Chilliwack Airport, Langley Central and Surrey East air quality monitoring stations (Figure 16d).  Hope Airport was the only station that experienced elevated ozone  concentrations on the  night of 6 January. After 6 January at 1000 U T C (0200  PST) until 8 January at 0400 U T C (7 January at 2000 PST), elevated ozone concentrations were measured at all of the rural air quality stations with the exception of Surrey East. These elevated ozone concentrations reached concentrations between 30-35 ppb. On 9 January at 0400 U T C (8 January at 2000 PST), the ground level ozone concentrations, with the exception of Langley Central decreased to near zero values before sharply increasing to 25-35 ppb on 9 January at 0600 U T C (8 January at 2200 PST). The ground level ozone concentrations remained elevated at the rural stations, with the exception of Hope Airport, for the rest of the episode. In the urban areas, the Kitsilano and Richmond South air quality monitoring stations exhibited a diurnal pattern with concentrations reaching 20-25 ppb on 6 January.  Downtown Vancouver experienced ground level ozone concentrations of  only 5-10 ppb during this time (Figure 16d).  For the rest of the episode, only  Richmond South experienced elevated ozone concentrations with maximum values  91  ranging between 15-20 ppb.  Downtown Vancouver and Kitsilano had no hourly  ground level ozone concentrations greater than 5 ppb during this time. During the episode, the A D M ozone concentration experienced two relative maxima and one relative minimum (Figure 15d).  The first relative A D M ozone  concentration maximum occurred on 5 January (33 ppb) and the second occurred on 9 January (29 ppb). The minimum occurred on 7 January (14 ppb). Associated with the relative maxima of A D M ozone concentration, the C V of ozone experienced deep minima of 0.15 ppb on 5 January and 0.2 ppb on 9 January. Conversely, the C V of ozone reached a maximum of 0.9 ppb on 7 January, the same day that the A D M ozone concentration was at its minimum. The shape of the daily B e activity concentration time series for the episode was 7  similar to that of Case #3, with low values at the beginning and end of the episode and a single spike at about 9000 uBqm" that occurred in the middle of the episode (8 3  January) (Figure 15d). Although no data was available at this time, any precipitation that might have occurred could explain the single sharp spike in the B e time series 7  for the episode (see sections 4.3.1.2,.4.3.2.2 and 4.3.3.2). The A D M temperature remained fairly constant during the episode with temperatures ranging between 7-10°C. 4.3.4.3  Back Trajectory Analysis  Starting on 6 January at 1800 U T C , trajectories arriving at 1000 m a.s.l. over the L F V originated at altitudes of between 6700 - 7500 m in a trough centered about 500 km south of the Aleutian Islands six days earlier (Figure 42). On 1 January at 0600 U T C , these trajectories experienced a rapid descent of between 2.9-5.8 cms" for the 1  next 24 hours. On January 2 at 0600 U T C , these trajectories began to ascend to 4500 m before descending into the L F V .  92  Time (hours)  Figure 42 - Three dimensional back trajectories arriving over the L F V area on 6 January 2001 at 1800 UTC using the CANERM model. The three day back trajectory ensemble consists of 27 trajectories starting at 100m a.s.l., 500m a.s.l. and 1000m a.s.l. over the LFV. The lower plot shows the time height profile of the trajectories. Back trajectory data courtesy of Environment Canada's Canadian Meteorological Center.  Between 6 January at 1200 U T C and 8 January at 0600 U T C , at least one third of the trajectories arriving over the L F V originated from altitudes greater than 5000 m. The most notable example occurred on 7 January at 0600 U T C when all but five of the trajectories originated above 5000 m six days earlier (Figure 43).  Trajectories  arriving over the L F V at an altitude of 500 m a.s.l. originated at heights ranging from 7000 m-9000 m in the Aelutian Island trough six days earlier. Between 1 January at 0600 U T C and 2 January at 0000 U T C , these trajectories experienced a rate of descent of between 1.7-7.0 cms" . Trajectories arriving over the L F V at an altitude of 1  1000 m a.s.l. on 7 January at 0600 U T C originated as far west as central Siberia six days earlier. Over this six day period, the trajectories traveled around the large scale  93  -144  -132  -120  -106  -83  -84  -72  -60  48  -36  -24  -12  0  Tirre (hours)  Figure 43 - Three dimensional back trajectories arriving over the L F V area on 7 January 2001 at 1800 UTC using the CANERM model. The three day back trajectory ensemble consists of 27 trajectories starting at 100m a.s.l., 500m a.s.l. and 1000m a.s.l. over the LFV. The lower plot shows the time height profile of the trajectories. Back trajectory data courtesy of Environment Canada's Canadian Meteorological Center.  Pacific trough at mid tropospheric levels between 1 January at 0600 U T C and 5 January 0000 U T C before descending into the L F V . The trajectories arriving over the L F V at 100 m a.s.l. originated in the Aleutian Island trough six days earlier at heights between 5000 m-6000 m. Because several of the trajectories originated from altitudes between 8000m-9000 m above sea level, some of the air over the L F V may have originated in the lower stratosphere six days prior to their arrival in the L F V . There are two reasons why this could be so.  First, the rapid descent experienced by some of the trajectories that  arrived over the L F V on 6 January at 1800 U T C and 7 January at 0600 U T C may have been due to the air descending into the troposphere via a tropopause fold.  94  Second, the locations of suspected tropopause folds can be identified by thin, dark, elongated lines in water vapor satellite images (Appenzeller and Davies, 1992, Stohl et al., 2000). These are possible areas of dry air descending from the stratosphere (Appenzeller and Davies, 1992). Figure 44 shows such a dry line that existed near the Alutian Island trough extending from the Kamchatka Peninsula to 3 0 ° N latitude 165°W longitude and then up to 4 0 ° N latitude 155°W longitude.  Figure 44 - GOES water vapor satellite image for 1 January 2001 at 0000 UTC. Dark regions denote large radiance values (low moisture content in the mid to upper troposphere). Image courtesy of Environment Canada's Canadian Meteorological Center.  95  Figure 45 - Possible stratospheric (dashed ellipses) and tropospheric (solid ellipses) source regions of ozone that arrived in the LFV during the four case studies in Section 4.3.  5. Discussion 5.1 Possible Source Regions of Ozone Using the trajectory figures in Section 4.3 (Figures 26, 30, 34-36 and 42-43), it is possible to draw some conclusions about the general non-local source regions of ozone in the L F V during the four case studies.  By non-local source regions, it is  meant that this ozone was not produced in-situ, but was transported from another regions by the synoptic and local wind. Figure 45 shows the possible stratospheric and tropospheric source regions of ozone that arrived in the L F V during the four case  96  studies in Section 4.3.  The largest possible source region of stratospheric ozone  encompasses an area between the central North Pacific Ocean and the Northern Arctic Ocean in the north south direction and Central Siberia and Eastern Alaska in the east west direction. With the exception of Figure 30 in Case Study 2 and Figure 34 in Case Study 3, all o f the trajectories originated from this source region at altitudes between 0 m a.s.l. to 9000 m a.s.l. It is important to note that this general region was also the location of many of the upper level troughs\cutoff lows and associated synoptic scale surface low pressure systems described in the case studies.  It was  suggested in Section 4.3 that these synoptic systems could have been responsible for the transport of ozone-rich stratospheric air into the troposphere by various mechanisms (i.e. tropopause folds, decent of stratospheric air into a cutoff low etc.). This region should therefore be the focus of future research into the long range transport of stratospheric ozone to the Pacific Northwest. Case Study 2 in Section 4.3 produced an interesting example where possibly both stratospheric and anthropogenically/naturally produced tropospheric ozone encountered along the same trajectory path (Figure 30).  were  During this case study  several trajectories originated over the eastern seaboard of the United States at altitudes around 2000 m. As these trajectories moved northwards, they could have encountered stratospheric air in a cutoff low situated over Northern Canada before traveling southwards and descending into the L F V .  In Section 4.3.2.3 it was noted  that similar multiple source trajectory case studies have been previously identified in the Swiss Alps by Stohl and Trickl (1999). These two possible source regions are indicated in Figure 45 as the circles situated over the eastern seaboard of the United States and over the Arctic region of Canada.  97  The final possible source region located over the western seaboard of the United States is tropospheric in origin.  This occurred during Case Study 3 (Figure 34).  Although all of the trajectories in this figure stayed below 1500 m a.g.l. during the entire back trajectory period, this example illustrates that, during an elevated ozone episode, the source regions could be either tropospheric (Figure 34) or stratospheric (Figures 35-36) in origin. It should be noted that, during the current research^ trajectories originating from lower tropospheric levels over the Eurasian continent  may be encountering  anthropogenically produced ozone that originated in this region.  This transport is  made possible by the warm conveyor belt (WCB) process whereby ozone and other anthropogenic pollutants are lofted into the upper troposphere by frontal systems, transported over a great distance over several days (in this case, across the Pacific Ocean) and then subsiding to the surface thereby elevating surface level pollutant concentrations.  Stohl et al. (2002) and Jaffe et al. (2003) have investigated W C B  case studies and performed modelling in the Eurasian and North Pacific.  5.2 General Trajectory Error — A Brief Discussion Although meteorological trajectories are used widely in atmospheric science, especially with research involving S T E (Stohl et al, 2002, Stohl and Trickl, 1999, Schuepbach et al, 1999, Sorensen and Nielsen, 2001, Viezee et al, 1983, Wakamatsu et al,  1989, Haagenson et al,  1981, K i m et al, 2002, Lamb, 1977, Parrish et al,  2000), the errors associated with trajectories are rarely reported or even acknowledged (Stohl et al, 2002).  The lack of acknowledgement of trajectory error has serious  consequences, especially if the computed trajectories are taken at face value with no validation against other meteorological data and analysis (Stohl et al, 2002). Rather than go into an in-depth trajectory error analysis, this section will cover some of the  98  sources of trajectory error and suggest some possible alternatives to simple single meteorological back trajectories. One of the larger sources of trajectory error can occur before the model is even run.  If the meteorological model output being fed into the trajectory model is  inaccurate, then the trajectory results will be inaccurate as well (Stohl, 1998). Even if the meteorological model output is of good quality, the grid point resolution of the output may be too coarse.  This requires a greater amount of interpolation between  grid points when the trajectory model is being run. When using trajectory models that utilize meteorological model output with a fine grid point resolution, one must be aware that the model being used to generate the meteorological model output in the first place may not have sufficient initialization data. This will thus require a fair amount of interpolation as that meteorological model is running. This could result in substantial interpolation errors in the meteorological model output. While the model is running, two main sources of error can occur. Firstly, error can occur when  the  wind velocity  is interpolated from grid points  in the  meteorological model output (Rolph and Draxler, 1990, Doty and Perkey, 1993, Stohl et al, 1995). This will lead to errors in the actual trajectory positions especially if the spacing between the grid points is large (Stohl et al, 2002). This error can be kept to a minimum if the quality of the meteorological model output is good and the spacing between the grid points is small.  A second error can occur when the numerical  solution to the trajectory equation is truncated due to computer memory and storage issues (Walmsley and Mailhot,  1983,  Seibert,  1993).  Because of the steady  improvement in numerical schemes, coupled with increasing computer memory and storage, truncation errors are very small when compared to interpolation errors (Stohl et al, 2002). When all of these factors are taken into account, the resulting trajectory  99  positional errors are typically on the order of 15%-20%, but can reach as high as 100% (Stohl et al, 2002). There  is  one  other  complication  that  must  be  considered  when  using  meteorological trajectories. This complication arises from the fact that trajectories are the paths of infinitesimally small particles of air and are thus not representative of finite volumes (Stohl et al, 2002). The reason for this is that finite volumes of air can be deformed by flowing around physical objects (i.e. flow around a mountain) or by wind shear in both the horizontal and vertical. Thus, a unit volume of air may be stretched by many units in the horizontal and the vertical as it transported by the flow (Seims et al, 2000, Pudykiewicz and Koziol, 1998). Because of this deformation, two infinitesimally small particles, initially close together in a finite volume, could be separated by large distances even over relatively short trajectory run times (Stohl et al, 2002). One possible solution to this problem would be to set up an ensemble of trajectories (Merrill et al,  1985, Kahl, 1993, Kahl, 1996). Using this technique, a  number of "supporting" trajectory end point receptors are set up around the main end point receptor so that a more representative sample of a unit finite volume may be realized.  5.3 CANERM and HYSPLIT Model Comparison During several measurement  campaigns involving transport of particulates,  pollutants and ionizing radiation, more than one trajectory model was often utilized (Hsu et al, 2003, Lupu and Maenhaut, 2002, N'dri Koffi et al, 1998, Baumann and Stohl, 1997). The major benefit of using multiple trajectory models is that it provides an opportunity for multiple agencies, institutions and governments to test and compare their model's output against other models and available atmospheric data such as balloon flight tracks (both manned and unmanned) (Baumann and Stohl,  100  1997). If large discrepancies occur between the models and/or between the available atmospheric data and the model output, the model can be corrected so that its output agrees more favorably with the available data.  One such experiment, E T E X  (European Tracer Experiment) involved the use of radiation and other data collected during and after the Chernobyl nuclear disaster in 1986.  Further details on this  project can be found in Graziani (1998) and Nodop (1997). Although a model intercomparison study like E T E X is far beyond the scope of the current research it is still necessary to compare the model output obtained from C A N E R M with model output from another model even if only at a basic level. The main objective was to determine which of the seven sets of ensemble back trajectories (Figures 26, 30, 34-36 and 42-43) in the case studies in Section 4.3 showed good positional agreement (i.e. latitude, longitude and height) between the two trajectory models and which ensemble back trajectory sets did not show good positional agreement. The trajectory model used for intercomparison was the H Y S P L I T model outlined in Section 4.2.1. In order to compare the positions of the trajectory paths produced by H Y S P L I T and C A N E R M , the H Y S P L I T model was first run during the same seven trajectory release times as C A N E R M indicated in Figures 26, 30, 34-36 and 42-43. Next, for each model run, the latitude, longitude and height of the nine trajectory paths originating at 100 m a.g.l., 500 m a.g.l. and 1000 m a.g.l. respectively were averaged every six hours.  This resulted in an average trajectory path for each end point  receptor height (100 m a.g.l., 500 m a.g.l. and 1000 m a.g.l.) for each model run. This procedure was carried out for both models. Finally, the positional (latitude, longitude and height) differences between these C A N E R M and H Y S P L I T average trajectory paths were found.  The resulting plots (Figure 46) show the absolute positional  101  differences between the two models over the entire back trajectory time for each of the seven model runs. Upon inspection of the plots in Figure 46, it is immediately apparent that there is a substantial amount of disagreement between the two models.  Generally, this  disagreement increases as the trajectory run progresses. The latitudinal disagreement was generally less than 10°, but differences as high as 18° did occur in several cases (Figures 46b and 46e). Typical longitudinal differences were almost twice as high at 20°, but maximum differences did climb to between 30° and 60°. Typically, the height differences were less than 1000 m for the first 24 - 72 hours. After 72 hours, the height differences rose quite sharply to values exceeding 3000 m in some cases. Only Figures 46c and 46f showed relatively small positional differences between the two models during the entire model run time with latitudinal (with one exception), longitudinal and height (with one exception) differences not exceeding 6°, 10° and 1200 m respectively. Figures 46a, 46b and 46e showed the least agreement between the two models in latitude, longitude and height. Although it is beyond the scope of the current research to speculate why these differences are so large, it is necessary to point out that care must be taken when performing trajectory studies. Trajectory errors and model disagreement can become quite large (Stohl et al, 2002) and supplemental atmospheric data (surface station data, ozone, radiation, synoptic analysis and satellite imagery) and analyses must be utilized to validate the trajectory results.  102  a).  -24  -4B  Trajectory  -12 time  -96  -120  -144  (hour)  Average C A N E R M values - Average HYSPLIT values for end point receptors at 100 m a.s.l. Average C A N E R M values - Average H Y S P L I T values for end point receptors at 500 m a.s.l. Average C A N E R M values - Average H Y S P L I T values for end point receptors at 1000 m a.s.l.  b).  -24  . -48  Trajectory  -72 time  -96  -120  (hour)  Average C A N E R M values - Average H Y S P L I T values for end point receptors at 100 m a.s.l. Average C A N E R M values - Average H Y S P L I T values for end point receptors at 500 m a.s.l. Average C A N E R M values - Average HYSPLIT values for end point receptors at 1000 m a.s.l.  103  c).  -24  -48  -72  Trajectory  time  -96  -24  (hour)  -48  Trajectory  -72 time  -96 (hour)  3500 3000 2500 2000 - | 1500  H  1000 500 - \  Trajectory  time  (hour)  Average C A N E R M values - Average HYSPLIT values for end point receptors at 100 m a.s.l. Average C A N E R M values - Average HYSPLIT values for end point receptors at 500 m a.s.l. Average C A N E R M values - Average HYSPLIT values for end point receptors at 1000 m a.s.l.  d).  -24  -48  Trajectory  -72 time  -96  -120  -144  (hour)  Average C A N E R M values - Average HYSPLIT values for end point receptors at 100 m a.s.l. Average C A N E R M values - Average HYSPLIT values for end point receptors at 500 m a.s.l. Average C A N E R M values - Average HYSPLIT values for end point receptors at 1000 m a.s.l.  104  e).  -24  -48  Trajectory  -72 time  -96  -120  -144  (hour)  Average C A N E R M values - Average HYSPLIT values for end point receptors at 100 m a.s.l. Average C A N E R M values - Average H Y S P L I T values for end point receptors at 500 m a.s.l. Average C A N E R M values - Average HYSPLIT values for end point receptors at 1000 m a.s.l.  -24  -12  -48  Trajectory  time  -96  -120  -144  -24  (hour)  3500 -1 a |  3  3000  -1  2500  "  x.  u oi ^  2000 ~  x  15 00  -48  Trajectory  -72 time  -96  -120  (hour)  1  -  -24  -48  Trajectory  -72 time  -96  -120  -144  (hour)  Average C A N E R M values - Average HYSPLIT values for end point receptors at 100 m a.s.l. Average C A N E R M values - Average HYSPLIT values for end point receptors at 500 m a.s.l. Average C A N E R M values - Average HYSPLIT values for end point receptors at 1000 m a.s.l.  105  g).  -24  -46  Trajectory  -72 time  -96  -120  -144  (hour)  Average C A N E R M values - Average HYSPLIT values for end point receptors at 100 m a.s.l. Average C A N E R M values - Average HYSPLIT values for end point receptors at 500 m a.s.l. Average C A N E R M values - Average HYSPLIT values for end point receptors at 1000 m a.s.l.  Figure 46 - Absolute latitudinal, longitudinal and height differences between the CANERM and HYSPLIT models for back trajectory runs starting at 100 m a.g.l (solid line), 500 m a.g.l. (dashed line) and 1000 m a.g.l. (stipled line) over the LFV for a) 31 October 1.  106  Figure 47 - Daily Be activity concentrations measured at the CTBTO IMS station CA002 (solid line) and total daily precipitation measured at Vancouver International Airport (dashed lines) during October 1996. 7  5.4 Precipitation and it Influence on Be Concentrations 7  In  Section 2.1  it was  noted that  7  B e was  susceptible  to scavenging by  precipitation. Figure 47 illustrates this phenomenon in two different ways.  Firstly,  when there are frequent precipitation events, the B e concentrations remain low. 7  Secondly, the daily B e activity concentration drops dramatically even during low to 7  moderate precipitation events. Figure 48 illustrates a period during which there is very little precipitation. Unlike Figure 47, a high B e concentration event can be seen in its 7  entirety with concentrations rising from background levels to its maximum and then slowly decreasing to background concentrations again. These examples illustrate the need for both wet and dry deposition measurements at radionuclide stations so that all of the high B e events may be captured. 7  107  10000  45  -  £  CT  CD 3.  g  - 35 8000 ~  0  —  I 02  1—'—I—1—I 04  06  08  1  1  1  1  1  10  12  14  16  18  August  1 — f 20  22  1  1  24  26  —  —  1  28  I 1 —  L  0  30  1997  Figure 48 - Daily 7Be activity concentrations measured at the CTBTO IMS station CA002 (solid line) and total daily precipitation measured at Vancouver International Airport (bars) during August 1997. In order to quantitatively determine whether precipitation reduces the daily B e 7  activity concentrations at the C T B T O IMS station CA002 on the U B C campus, a simple one sided t-test was performed. For this t-test, the null hypothesis was that the average daily B e activity concentration during days in which there was measured 7  precipitation was equal to the average daily B e activity concentration during days in 7  which there was no measured precipitation and that any variation was due strictly to chance.  The alternative hypothesis  was that the average daily  7  B e activity  concentration during days in which there was measured precipitation was significantly lower than the average daily B e activity concentration during days in which there 7  was no measured precipitation. The standard choice of significance level (or P value) for a one sided test of 2.5% (0.025) was used to accept or reject the null hypothesis (Samuels, 1989). Using a one sided t-test and an unpooled standard error, a P value  108  of 2.5x10" was found. Since this P value is much less than the chosen significance 7  level of 0.025, the null hypothesis was rejected in favor of the alternative hypothesis. This means that the difference between the average daily B e activity concentration 7  during days in which there was precipitation was significantly lower than days in which there was no precipitation. Thus precipitation does reduce the dry deposition 7  Be activity concentrations at the C T B T O IMS station CA002 at U B C .  6. Conclusions Using hourly ozone and daily B e activity concentration data during the period of 7  10 April 1996 - 12 March 2002, two different detection algorithms uncovered several episodes in which stratospheric ozone may be influencing ground level ozone concentrations in the L F V . The 80  percentile detection algorithm was more  th  effective in uncovering spring and summer episodes because the criteria behind this algorithm depended on both high daily B e activity concentrations and high A D M 7  ozone concentrations. Because the A D M ozone concentration was required to be in the 80  th  percentile, a large number of episodes in which anthropogenically derived  ozone may have had a strong influence on ground level ozone concentrations may have been included in these episodes.  This argument was supported by Figure 14  which showed that during the months in which the 80  th  percentile algorithm was  selecting episodes, the A D M temperature was high and would thus tend to produce favorable conditions for anthropogenically derived ozone production. The concern about these episodes was that the anthropogenically derived ozone concentration may be so large that it effectively drowns out the stratospheric contribution to the total ground level ozone concentration. In order to partially alleviate this problem, months during which the A D M temperatures were high were eliminated from the data set before the 80 percentile algorithm was applied. Although a few episodes during late th  109  fall and late winter were uncovered (Figures 20-21), the algorithm failed to uncover any episodes between November and January. The second algorithm relied only on high B e activity concentrations in the hope 7  that more episodes would be uncovered during the fall and winter months when the A D M temperatures were low and thus stratospheric ozone would be making a greater contribution to the total ground level ozone concentration. The resulting episodes that were detected by this algorithm took place during all months of the year with several episodes occurring during the late fall and winter months.  Although the ozone  concentrations were low during this time, a number of interesting episodes were uncovered during which the hourly ozone concentrations, especially in the rural areas, exhibited a classical stratospheric pattern whereby ozone concentrations rose sharply from near zero background levels to values exceeding 30 ppb within the span of only a few hours. These elevated concentrations then remained fairly constant for several days (with very little or no nocturnal N O scavenging) before slowly decreasing to x  background levels again. Four episodes of the type described above were more fully investigated using meteorological back trajectories, surface temperature and precipitation measurements, synoptic charts and satellite data. The goals of these case studies were to uncover possible source regions of stratospheric ozone and the transport pathways of this ozone from these source regions to the L F V .  During these case studies, a large  number of back trajectories arriving over the L F V were found to have originated over an area spanning Eastern Russia and China, Japan, the North Pacific and Arctic Oceans and Alaska (Figure 45). Many trajectories within this region reached altitudes between 5000 m to over 9000 m a.g.l. During these case studies, a large number of upper level troughs/cutoff lows and surface low pressure systems were present in this  110  area.  Pervious research has indicated that these synoptic systems are largely  responsible for the injection of ozone-rich stratospheric air into the troposphere by way of tropopause folds.  These case studies also uncovered situations in which  multiple ozone sources were intercepted by the same trajectory. Using a simple one sided t-test, it was found that precipitation does decrease the daily B e dry deposition activity concentrations. 7  This would suggest that wet  deposition measurements of B e should be taken although the cost may limit these 7  types of measurements to case studies only. Although a numerical analysis of the absolute contribution of stratospheric ozone to the total ground level ozone concentration is beyond the scope of this research, it is important to note that, during the case studies described in Section 4, two features become apparent. Firstly, the elevated ozone concentrations during the case studies seem to be making an important contribution in the rural areas. Using Figure 16 as a rough guide, a total stratospheric contribution of up to 10-20 ppb may be considered to be a reasonable first guess. Secondly, there is very little nocturnal N O scavenging x  in the rural areas during these episodes. The total nocturnal stratospheric contribution will therefore be greater than during the daytime. Again, using Figure 16 as a rough guide, the total nocturnal stratospheric contribution could be as high as 20-40 ppb. This stratospheric contribution can thus not be ignored when decisions are being made with regards to ground level ozone reduction programs.  It addition, the current  research raises the question about the health effects resulting from long term (on the order of a few days) exposure to constant moderate ozone concentrations as experienced in the rural areas in Figure 16 as opposed to acute exposure (on the order of a few hours) to high ozone concentrations. future research.  Ill  This may be an interesting topic of  6.1 Future Research In the future, the current research may be extended to include the following: 1. Using the methods outlined in the current research, B e and ozone data could be 7  made available from other areas of Canada so that the source regions of stratospheric ozone could be determined for those areas. 2.  The synoptic types (McKendry 1994, Yarnal 1993) under which stratospheric ozone is detected in the L F V should be investigated using existing synoptic typing techniques.  It may be that a sequence of synoptic types will have to be defined  since S T E that leads to increased surface concentrations of ozone and B e in the 7  L F V is suspected to occur far offshore. 3. During periods in which these S T E synoptic types are forecasted to occur, wet deposition measurements could be taken at the C T B T O IMS station CA002 at U B C so that the total B e and ozone transported from the stratosphere to the 7  surface may be measured. 4.  Greater use of water vapor satellite data could be used in conjunction with a long range atmospheric dispersion model to determine the spatial structure and fate of stratospheric air that is injected into the troposphere. This technique has already been utilized by Stohl et al. (2000). agreement  between dry lines  Briefly, Stohl et al. (2000) found good  in satellite  imagery (possible  areas of dry  stratospheric air descending into the troposphere) and high ozone tracer mixing ratios between 2000 m and 4000 m a.s.l. predicted by the dispersion model. 5. In addition to B e , 7  10  B e (1/2 life = 1.5 x 10 years) measurements could be taken 6  and a B e / B e ratio could thus be utilized. Because dry and wet deposition are 10  7  negligible in the stratosphere and the half life of B e is so much larger than Be, 10  10  Be/ Be 7  ratios are normally higher in the stratosphere  112  7  compared to the  troposphere. Like B e , B e also attaches to aerosols which are then susceptible to 7  scavenging. necessary.  10  Wet deposition measurements of either species are therefore not Thus, at the surface, high dry deposition  possible S T E episodes.  10  B e / B e ratios denote 7  Unfortunately B e has to be measured using accelerator 10  mass spectrometry (compared to high-resolution gamma spectrometry for Be), 7  which can be quite expensive.  So far, only short term studies such as those  outlined in Zanis et al (2003) and Dibb et al. (1994).  113  7. References Ancellet, G . , Pelon, J., Beekman, M . , Papayannis, A . , Megie, G . , 1991: Ground-based lidar studies of ozone exchanges between the stratosphere and the troposphere. Journal of Geophysical Research, 96, 22401 -22421. Appenzeller, C , Davies, H . C . , 1992: Structure of stratospheric intrusions into the troposphere. Nature, Vol. 358, 570-572. Bates, D.R., Nicolet, M . , 1950: The photochemistry of atmospheric water vapour.  Journal of Geophysical Research, 55, 301. Baumann, K . , Stohl, A . , 1997: Validation of a long-range trajectory model using gas balloon tracks from the Gordon Bennett Cup 95. Journal of Applied  Meteorology, 36, 711-720. Bonasoni, P., Evangelisti, F., Bonafe, U . , Ravegnani, F., Calzolari, F., Stohl, A . , Tositti, L . , Tubertini, O., Colombo, T., 1999: Stratospheric ozone intrusion episodes recorded at Mt. Cimone during the V O T A L P project: case studies.  Atmospheric Environment, 34, 1355-1365. Bovis, P., Steyn, D . G . , 2001: Stratosphere-troposphere exchange and its influence on surface ozone concentrations in the Lower Fraser Valley - A preliminary report. Report published internally to Health Canada. Caillet, S., Arpagaus, P., Monna, F., Dominik, J., 2001: Factors controlling B e and P b atmospheric deposition as revealed by sampling individual rain events in the region of Geneva, Switzerland. Journal of Environmental Radioactivity, 53, 241-256. 7  2 , 0  Chapman, S.A., 1930: A theory of upper atmospheric ozone.  Royal Meteorological  Society, 3, 103. Cho, J.Y.N., Newwll, R . E . , Bui, T.P., Browell, E . V . , Fenn, M . A . , Mahoney, M . J . , Gregory, G . L . , Sachse, G . W . , Vay, S.A., Kucsera, T . L . , Thompson, A . M . , 1999: Observations of convective and dynamical instabilities in tropopause folds and their contribution to stratosphere-troposphere exchange. Journal of  Geophysical Research, 104 (D17), 21,549-21,568. Chung, Y.S., Dann, T . , 1984: Observations of stratospheric ozone at the ground level in Regina, Canada. Atmospheric Environment, 19, 157-162. Danielson, E . F . , 1968: Stratospheric-tropospheric exchange based on radioactivity, ozone and potential vorticity. Journal of the Atmospheric Sciences, 52, 502518. Danielsen, E . , Bleck, R., Shedlovsky, J., Wartburg, A . , Haagenson, P., Pollock, W., 1970: Observed distribution of radioactivity, ozone and potential vorticity associated with tropopause folding. Journal of Geophysical Research, 75, 2353-2361.  114  Danielsen, E . F . , Mohnen, V . A . , 1977: Project Dustorm report: ozone transport, in situ measurements and meteorological analysis of tropopause folding. Journal of  Geophysical Research, 82, 5867-5877. Davies, T . D . , Schuebach, E . , 1993: Episodes of high ozone concentrations at the earth's surface resulting from transport down from the upper troposphere/lower stratosphere: review and case studies. Atmospheric Environment, 28, 53-68. Dibb, J.E., Meeker, L . D . , Finkel, R . C . , Southon, J.R., Caffee, M . W . , Barrie, L . A . , 1994: Estimation of stratospheric input into the Arctic troposphere: B e and B e in aerosols in Alert, Canada. Journal of Geophysical Research, 99, 12855-12864. 7  10  Doty, K . G . , Perkey, D.J., 1993: Sensitivity of trajectory calculations to the temporal frequency of wind data. Monthly Weather Review, 121, 387-401. Draxler, R.R., Hess, G . D . , 1997: Description of the Hysplit_4 modeling system.  NOAA Tech Memo ERL ARL-224, Dec, 24p. Dutkiewicz, V . A . , Husain, L . , 1985: Stratospheric and tropospheric components of B e in surface air. Journal of Geophysical Research, 90, 5783-5788. 7  Fast, J.D., Berkowitz, C M . , 1997: Evaluation of back trajectories associated with ozone transport during the 1993 North Atlnatic Regiona Experiment.  Atmospheric Environment, 31, 825-837. Feely, H . , Larsen, R., Sanderson, C , 1989: Factors that cause seasonal variations in beryllium-7 concentrations in surface air. Journal of Environmental  Radioactivity, 9, 223-249. Gislason, K . B . , Prahm, L . P . , 1983: Sensitivity study of air trajectory long-range  transport modelling. Atmospheric Environment, 12, 2463-2472. Graziani, G . , Galmarini, S., Grippa, G . , Klug, W . (Eds.), 1998: Real-Time LongRange Dispersion Model Evaluation of the Second E T E X Release. Luxembourg, Office for Official Publications of the European Communities. Gusten, H . , Heinrirh, G . , Cvitas, T., Klasing, L . , Ruscic, B., Lalas, P.D., Petrakis, M . , 1988: Photochemical formation and transport of ozone in Athens, Greece.  Atmospheric Environment, 22, 1855-1861. Haagenson, P.L., Shapiro, M . A . , Middleton, P., Laird, A . R . , 1981: A case study relating high ground level ozone to enhanced photochemistry and isentropic transport from the stratosphere. Journal of Geophysical Research, 86, 52315237. Haagen-Smit, A . J . , 1952: Chemistry and physiology of Los Angeles smog. Engineering Chemistry, 44, 1342-1346.  Industrial  Haagen-Smit, A . J . , Bradley, C . E . , Fox, M . M . , 1953: Ozone formation in chemical oxidation of organic substances. Industrial Engineering Chemistry, 45, 2086-.  115  Haagen-Smit, A . J . , Darley, E . F . , Zaitlin, M , Hull, H . , Noble, W., 1951: Investigation on injury to plants from air pollution in the Los Angeles area. Plant  Physiology, 27, 18-. Haagen-Smit, A . J . , Fox, M . M . , 1954: Photochemical hydrocarbons and automobile exhaust! Journal of  ozone  formation  with  the Air Pollution Control  Association, 4, 105-109. Haagen-Smit, A . J . , Fox, M . M . , 1955: Automobile exhaust and ozone formation. S A E Technical Series No. 550277. Warrendale, Penn.: Society of Automotive Engineers. Haagen-Smit, A . J . , Fox, M . M . , 1956: Ozone formation in photochemical oxidation of organic substances. Industrial Engineering Chemistry, 28, 1484-. Heffter, J.L., Stunder, B.J.B., 1993: Volcanic Ash Forecast Transport And Dispersion (V A F TA D ) Model. Weather Forecasting, 8, 534-541. Hotzl, H . , Winkler, R., 1987: Activity concentrations of R a , R a , P b , K and B e and their temporal variations in surface air. Jounal of Environmental 2 2 6  2 2 8  2 , 0  4 0  7  Radioactivity, 5, 445-458. Hsu, T - K . , Holsen, T . M . , Hopke, P.K., 2003: Comparison of hybrid receptor models to locate P C B sources in Chicago. Atmospheric Environment, 37, 545-562. Ishikawa, Y . , Murakami, H . , Sekine, T., Yoshihara, K . , 1995: Precipitation scavenging studies of radionuclides in air using B e . Journal of Environmental Radioactivity, 26, 19-36. 7  Janach,  W . E . , 1989: Surface ozone: trend details, seasonal variations, and interpretation. Journal of Geophysical Research, 94, 18289-18295.  Johnson, W . B . , Viezee, W . , 1980: Stratospheric ozone in the troposphere - I. Presentation and interpretation of aircraft measurements. Atmospheric  Environment, 15, 1309-1323. Juarez, A . , Gray, C , Bravo, J.L., 1994: Influence of urban ozone in the measurements of the total ozone column in Mexico City. ATMOSFERA, 8, 35-43. Kahl, J.D., 1993: A cautionary note on the use of air trajectories in interpreting atmospheric chemistry measurements. Atmospheric Environment, 21A, 30373038. Kahl, J.D., 1993: A cautionary note on the use of air trajectories in interpreting atmospheric chemistry measurements. Atmospheric Environment, 2 1 A, 30373038. Kato, H . , Fujita, S-I., Nishinomiya, S., 1990: Mechanism of spring high-oxidant episode - a meteorological analysis in and around the Hokuriku District, Japan. Atmospheric Environment, 24A, 2023-2033.  116  Kim, Y . K . , Lee, H . W . , Park, J.K., Moon, Y . S . , 2002: The stratosphere-troposphere exchange of ozone and aerosols over Korea. Atmospheric Environment, 36, 449-463. Lai, D., Peters, B., 1967: Cosmic ray produced radioactivity on the earth. Handbuch  derPhysik, 46, 551-612. Lamb, R . G . , 1977: A case study of stratospheric ozone affecting ground level oxidant concentrations. Journal ofApplied Meteorology, 16, 780-794. Lupu, A . , Maenhaut, W., 2002: Application and comparison of two statistical trajectory techniques for identification of source regions of atmospheric aerosol species. Atmospheric Environment, 36, 5607-5618. Mayer, H . , 1999: Air Pollution in cities.  Atmospheric Environment, 33, 4029-4037.  McKendry, I.G., 1994: Synoptic circulation and summertime ground-level ozone concentrations at Vancouver, British Columbia. Journal of Applied  Meteorology, 33, 627-941. Measday, D . F . , Stocki, T . J . , Mason, L . R . , Williams, D . L . , 2001: Detection of anthropogenic radionuclides by the CA002 monitoring station for the comprehensive test ban treaty. Health Physics, 80, 126-136. Medici, F., The IMS radionuclide network of the C T B T .  Radiation Physics and  Chemistry, 61, 689-690. Megumi, K . , Matsunami, T . , Ito, M . , Kiyoda, S., Mizohata, A . , Takeyoshi, A . , 2000: Factors, especially suinspot number, causing variations in surface air  concentrations and depositions of Be in Osaka, Japan. Geophysical Research Letters, 27, 361-364. 7  Merrill, J.T., Bleck, R., Avila, L . , 1985: modeling atmospheric transport to the Marshall Islands. Journal of Geophysical Research, 90, 12927-12936. Miller, C , Filkin, D . L . , Owens, A . J . , Steed, J . M . , Jesson, J.P., 1981: A two dimensional model of stratospheric chemistry and transport. Journal of  Geophysical Research, 86, 1379. Morison, R.P., Leslie, L . M . , Speer, M.S., 2002: Atmospheric modeling of air pollution as a tool for environmental prediction and management.  Meteorology and Atmospheric Physics, 80, 141-151.  Murao, N . , Ohta, S., Furuhashi, N . , Mizoguchi, I., 1990: The causes of elevated  concentrations of ozone in Sapporo. Atmospheric Environment, 24A, 15011507. National Research Council, 1992: Rethinking the ozone problem in urban and regional air pollution. Washington, National Academy Press.  117  N'dri Koffi, E . , Nodop, K . , Benech, B . , 1998: Comparison of constant volume balloons, model trajectories and tracer transport during E T E X . Atmospheric  Environment, 34, 4139-4149. Nodop, K . (Ed.), 1997: Proceedings of the E T E X Symposium on Long-Range Atmospheric Transport, Model Verification and Emergency Response. Luxembourg, Office for Official Publications of the European Communities. Paatero, Jussi and Juha Hatakka, 2000: Source areas of airborne B e and 7  2 1 0  Pb  measured in Northern Finland. Health Physics, 76, 691-696. Parrish, D . D . , Holloway, J.S., Jakoubek, R., Trainer, M . , Ryerson, T . B . , Hubler, G . , Fehsenfeld, F . C . , Moody, J.L., Cooper, O.R., 2000: Mixing of anthropogenic pollution with stratospheric ozone: a case study from the North Atlantic wintertime troposphere. Journal of Geophysical Research, 105, 24363-24374. Pont, V . , Fontan, J . , 2000: Local and regional contributions to photochemical atmospheric pollution in southern France. Atmospheric Environment, 34, 5209-5223. Prinn, R . G . , Alyea, F . M . , Cunnold, D . M . , 1978: Photochemistry and dynamics of the ozone layer. Annual Review of the Earth and Planetary Sciences, 6, 43. Pudykiewicz, J., 1988 : Numerical simulation of the transport of radioactive cloud from the Chernobyl nuclear accident. Tellus, 40B, 241-259. Pudykiewicz, J., 1989 : Simulation O f The Chernobyl Dispersion With a 3-D Hemispheric Tracer Model. Tellus, 41B, 391-412. Pudykiewicz, J.A., Koziol, A . S . , 1998: A n application of the theory of kinematics of mixing to the study of tropospheric dispersion. Atmospheric Environment, 32, 4227-4244. Rao, S.T., Sistla, G . , Schere, K . , Godowitch, J., 1991: Analysis of ozone air quality over the New York metropolitan area. A i r Pollution Modeling and its Application, V o l . VIII. Plenum Press, New York, N Y , 111-121. Realtime Environmental Applications and Display System ( R E A D Y ) website. N O A A Air Resource Laboratory (http://www.arl.noaa.gov/ready.html). Rolph, G . D . , Draxler, R.R., 1990: Sensitivity of three dimensional trajectories to the spatial and temporal densities of the wind field. Journal of Applied  Meteorology, 29, 1043-1054. Samuels, M . L . , 1989: Statistics for the life sciences. Englewood Cliffs, Prentice Hall. Schuepbach, E . , Davies, T . D . , Massacand, A . C . , 1998: A n unusual springtime ozone episode at high elevation in the Swiss Alps: contributions from both crosstropopause exchange and from the boundary layer. Atmospheric Environment, 33, 1735-1744.  118  Seibert, P., 1993: Convergence and accuracy of numerical methods for trajectory calculations. Journal ofApplied Meteorology, 32, 558-566. Seims, S.T., Hess, G . D . , Suhre, K . , Businger, S., Draxler, R., 2000: The impact of wind shear on observed and simulated trajectories during the A C E - 1 Lagrangian experiments. Australian Meteorological Magazine, 30, 579-587. Seinfeld, J.H., 1986: Atmospheric chemistry and physics of air pollution. New York, John Wiley & Sons, Inc. Sorensen, J . H . , Nielsen, N . W . , 2001: Intrusion of stratospheric ozone to the free troposphere through tropopause folds - a case study. Physical Chemistry of  the Earth (B), 26, 801-806. Stewart, R.W., Hameed, S., Pinto, J.P., 1977: Photochemistry of tropospheric ozone. Journal of Geophysical Research, 82, 3134-. Steyn, D . G . , Bottenhiem, J.W., Thompson, R . B . , 1997: Overview of tropospheric ozone in the Lower Fraser Valley, and the Pacific '93 field study.  Atmospheric Environment, 31, 2025-2035. Stohl, A . , 1998: Computation, accuracy and applications of trajectories - a review and  bibliography. Atmospheric Environment, 32, 947-966. Stohl, A . , Eckhardt, S., Forster, C , James, P., Spichtinger, N . , Seibert, P., 2002: A replacement for simple back trajectory calculations in the interpretation of atmospheric trace substance measurements. Atmospheric Environment, 36, 4635-4648. Stohl, A . , Seibert, P., 1998: Accuracy of trajectories as determined from the conservation of meteorological tracers. Quarterly Journal of the Royal Meteorological Society, 125, 1465-1484. Stohl, A . , Spichtinger-Rakowsky, N . , Bonasoni, P., Feldmann, H . , Memmesheimert, M . , Scheel, H . E . , Trickl, T., Hubener, S., Ringer, W., Mandl, M . , 1999: The influence of stratospheric intrusions on alpine ozone concentrations.  Atmospheric Environment, 34, 1323-1354. Stohl, A . , Trickl, T . , 1999: A textbook example of long-range transport: Simultaneous observation of ozone maxima of stratospheric and North American origin in the free troposphere over Europe. Journal of Geophysical Research, 104  (D23), 30,445-30-462. Stohl, A . , Wotawa, G . , 1995: A method for computing single trajectories representing  boundary layer transport. Atmospheric Environment, 29, 3235-3239. Stohl, A . , Wotawa, G . , Seibert, P., Kromp-Kolb, H . , 1995: Interpolation errors in wind fields as a function of spatial and temporal resolution and their impact on different types of kinematic trajectories. Journal ofApplied Meteorology, 34, 2149-2165.  119  Talpos, S., Cuculeanu, V . , 1997: A study of the vertical diffusion of B e in the atmosphere. Journal of Environmental Radioactivity, 36, 93-106. 7  Tremblay, J., Servranckx, R., 1992: Beryllium-7 as a case of stratospheric ozone: a case study. Journal of Radioanalytic and Nuclear Chemistry, 172,49-56. Tulet, P., Crassier, V . , Rosset, R., 2000: Air pollution modelling at a regional scale.  Environmental and Modelling Software, 15, 693-701. Viezee, W . , Johnson, W . B . , Singh, H.B., 1983: Stratospheric ozone in the lower troposphere II - assessment of downward flux and ground-level impact.  Atmospheric Environment, 17, 1979-1993. Wakamatsu, S., Uno, I., Ueda, H . , Uehara, Tateishi, K . H . , 1989: Observational study of stratospheric ozone intrusions into the lower troposphere. Atmospheric  Environment, 23, 1815-1826. Walmsley, J.L., Mailhot, J., 1983: On the numerical accuracy of trajectory models for long-range transport of atmospheric pollutants. Atmospheric Ocean, 21, 1439. Wayne, R.P., 1999: Chemistry of atmospheres. New York, Oxford University Press. Wunderli, S., Gehrig, R., 1991: Influence of temperature on formation and stability of of surface P A N and ozone. A two year field study in Switzerland.  Atmospheric Environment, 25A, 1599-1608. Yarnal, B . , 1993: Synoptic climatology in environmental analysis: a primer. Boca Raton, Belhaven Press. Zanis, P., Gerasopoulos, E . , Priller, A . , Schnabel, C , Stohl, A . , Zerefos, C , Gaggeler, H.W., Tobler, L . , Kubik, P.W., Kanter, H.J., Scheel, H . E . , Luterbacher, J., Berger, M . , 2003: A n estimate of the impact of stratosphere-to-troposphere transport (STT) on the lower free tropospheric ozone over the Alps using B e 10  and Be measurements. Atmospheric Environment. 108. Zanis, P., Schuepbach, E . , Gaggler, H.W., Hubener, S., Tobler, L . , 1999: Factors controlling B e at Jungfraujoch in Switzerland. Tellus, 51B, 789-805. 7  Zannetti, P., 1990: A i r pollution modeling: theories, computational methods and available software. Boston, Computational Mechanics Publications.  120  

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.831.1-0052336/manifest

Comment

Related Items